International Journal of Intelligent Systems and Applications in Engineering
https://ijisae.org/index.php/IJISAE
<div style="border: 3px solid black; padding: 10px; background-color: aliceblue;"> <p style="margin: 5px; font-size: 15px;"><strong style="font-size: 20px;"><u>Update Regarding Article's Indexing:</u></strong><br />Dear esteemed authors and readers,<br />We are pleased to inform you that the <strong>International Journal of Intelligent Systems and Applications in Engineering (IJISAE)</strong> has successfully passed the re-evaluation process by <strong>Elsevier</strong>. This achievement reflects our commitment to maintaining the highest standards of quality in academic publishing.<br />We are also excited to announce that our pending articles will start getting indexed in Scopus in 6 weeks. This is a significant milestone for us, and we believe it will enhance the visibility and accessibility of our published research.<br />We would like to express our gratitude to all our authors, reviewers, and readers for their continuous support and contributions towards making IJISAE a leading platform for scholarly research in the field of intelligent systems and applications in engineering.<br />We look forward to continuing to provide a high-quality platform for academic exchange and encourage all interested authors to submit their best work to IJISAE.<br /><br />Best regards,<br />The IJISAE Editorial Team</p> <br /> <p style="margin: 5px; font-size: 15px;"><strong style="font-size: 20px;"><u>Information for Authors:</u></strong><br />We are pleased to inform that we are now collaborating with <strong>Digital Commons, Elsevier</strong> for much better visibility of journal. Further authors will be able to observe their citations, metric like PlumX from journal website itself. <strong>IJISAE</strong> will be in transition from <strong>OJS</strong> to <strong>Digital Commons Platform</strong> in next few months so if their is any queries or delays contact directly on <em><strong>editor@ijisae.org</strong></em></p> </div> <p><strong><a href="https://ijisae.org/IJISAE">International Journal of Intelligent Systems and Applications in Engineering (IJISAE)</a></strong> is an international and interdisciplinary journal for both invited and contributed peer reviewed articles that intelligent systems and applications in engineering at all levels. The journal publishes a broad range of papers covering theory and practice in order to facilitate future efforts of individuals and groups involved in the field. <strong>IJISAE</strong>, a peer-reviewed double-blind refereed journal, publishes original papers featuring innovative and practical technologies related to the design and development of intelligent systems in engineering. Its coverage also includes papers on intelligent systems applications in areas such as nanotechnology, renewable energy, medicine engineering, Aeronautics and Astronautics, mechatronics, industrial manufacturing, bioengineering, agriculture, services, intelligence based automation and appliances, medical robots and robotic rehabilitations, space exploration and etc.</p> <p>As an Open Access Journal, IJISAE devotes itself to promoting scholarship in intelligent systems and applications in all fields of engineering and to speeding up the publication cycle thereof. Researchers worldwide will have full access to all the articles published online and be able to download them with zero subscription fees. Moreover, the influence of your research will rapidly expand once you become an Open Access (OA) author, because an OA article has more chances to be used and cited than does one that plods through the subscription barriers of traditional publishing model.</p> <p><strong>IJISAE (ISSN: 2147-6799)</strong> indexed by <a href="https://www.scopus.com/sourceid/21101021990#tabs=0" target="_blank" rel="noopener">SCOPUS</a>, <a href="https://app.trdizin.gov.tr/dergi/TVRBM05UVT0/international-journal-of-intelligent-systems-and-applications-in-engineering" target="_blank" rel="noopener">TR Index</a>, <a href="https://journals.indexcopernicus.com/search/details?jmlId=3705&org=International%20Journal%20of%20Intelligent%20Systems%20and%20Applications%20in%20Engineering,p3705,3.html">IndexCopernicus</a>, <a href="http://globalimpactfactor.com/intelligent-systems-and-applications-in-engineering-ijisae/%20in%20Engineering,p3705,3.html" target="_blank" rel="noopener">Global Impact Factor</a>, <a href="http://cosmosimpactfactor.com/page/journals_details/6400.html" target="_blank" rel="noopener">Cosmos</a>, <a href="https://scholar.google.com.tr/scholar?q=IJISAE&btnG=&hl=tr&as_sdt=0%2C5">Google Scholar</a>, <a href="http://www.journaltocs.ac.uk/index.php?action=search&subAction=hits&journalID=29745" target="_blank" rel="noopener">JournalTocs</a>, <a href="https://www.idealonline.com.tr/IdealOnline/lookAtPublications/journalDetail.xhtml?uId=679" target="_blank" rel="noopener">IdealOnline</a>, <a href="http://oaji.net/journal-detail.html?number=5475" target="_blank" rel="noopener">OAJI</a>, <a href="https://www.researchgate.net/journal/International-Journal-of-Intelligent-Systems-and-Applications-in-Engineering-2147-6799" target="_blank" rel="noopener">ResearchGate</a>, <a href="http://esjindex.org/search.php?id=2455" target="_blank" rel="noopener">ESJI</a>, <a href="https://search.crossref.org/" target="_blank" rel="noopener">Crossref</a>, and <a href="https://portal.issn.org/resource/ISSN/2147-6799" target="_blank" rel="noopener">ROAD</a>.</p> <p>Please Contact: <a href="mailto:editor@ijisae.org">editor@ijisae.org</a></p> <p><img style="width: 36px; height: 36px;" src="https://ijisae.org/public/site/images/ilkerozkan/about-the-author-1.jpg" alt="" align="left" /></p> <p><strong>Submit your manuscripts </strong><a style="color: blue;" href="http://manuscriptsubmission.net/ijisae/index.php/submission/about/submissions#authorGuidelines">Detail information for authors </a></p> <p><strong>Publication Fee:</strong> 600 USD (The APC is calculated based on the number of pages and color figures per page of the final accepted manuscript. Charges are fix 600 USD for first 10 pages. For manuscripts exceeding 10 pages, there will be an additional charge of USD 95 per additional page.)</p>en-USInternational Journal of Intelligent Systems and Applications in Engineering2147-6799<p>All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.</p> <p>IJISAE open access articles are licensed under a <a href="http://creativecommons.org/licenses/by-sa/4.0/" target="_blank" rel="noopener">Creative Commons Attribution-ShareAlike 4.0 International License</a>. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.</p>A Comparative Analysis of ARIMA and VAR Algorithms for Performance Analysis of High-Speed Diesel Pumps
https://ijisae.org/index.php/IJISAE/article/view/4832
<p>The demand for precise and efficient forecasting of High-Speed Diesel (HSD)pump performance is critical for optimizing fuel distribution, operational planning, and resource allocation in the petroleum industry. This paper presents a comprehensive comparison analysis of implementing two widely used time series forecasting algorithms, Auto regressive Integrated Moving Average (ARIMA) and Vector Auto Regression (VAR), for predicting vibration in electrical systems. The study spans a year-long dataset collected at various intervals, including seconds, minutes, hours, days, weeks, months, and yearly intervals, leveraging data from voltage, current, and temperature sensors. The research analyzes "Mean Squared Error (MSE), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE)" as three critical indicators for evaluating how well ARIMA and VAR perform. The analysis reveals that ARIMA consistently outperforms VAR across all intervals, demonstrating superior accuracy in predicting vibration levels. The data The dataset collected from a range of sensors provides a diverse and rich source of information, effectively capturing the electrical system's dynamic behavior. The results highlight the significance of selecting an appropriate forecasting model for time series data, especially system reliability and maintenance applications. This research contributes to the ongoing discourse on algorithm selection in time series forecasting for electrical systems and provides valuable insights for practitioners and researchers alike. The findings underscore the importance of considering the dataset's specific characteristics and the nature of the target variable when choosing between ARIMA and VAR algorithms for predictive modeling.</p>Smita MahajanShivali Amit WagleNihar RanjanSantosh Borde
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s0113Performance Analysis of Transfer Learning Framework for the Detection of Polyps in Colorectal Cancer
https://ijisae.org/index.php/IJISAE/article/view/4833
<p>Colorectal cancer (CRC) begins in the colon or rectum, gastrointestinal tract organs. It is a common cancer that causes many cancer deaths worldwide. CRC usually starts with a polyp, a benign growth that can become cancerous. CRC prevention, treatment, and control require early detection and treatment. In this study, we reviewed various, pertinent research based on CRC diagnostic techniques, colonoscopy, and the use of AI screening. We performed various quantitative and qualitative comparative analyses of diagnostic techniques based on numerous features. Colonoscopy and sigmoidoscopy allow doctors to examine the colon and rectum for abnormalities. Deep learning (DL) techniques in medical imaging and Artificial Intelligence (AI) have improved CRC diagnosis, particularly polyp detection. We discussed the present and possible use of AI, DL in CRC diagnosis. A sigmoidoscopy, a minimally invasive procedure, shows the potential in terms of reducing the number of incidences and mortality. Colonoscopy was the most invasive technique and possesses the risk of morbidity. The Markov model demonstrated that cost per life can be saved for a colonoscopy performed once in 10 years. Thus, colonoscopy certainly proves to be a golden standard with highest sensitivity with the capability of biopsy during diagnosis. The proposed pre-trained VGG19 model confirmed 97% accuracy in polyp detection when applied with the approach of Transfer Learning (TL). The model is not overfitting and is proven to be more accurate than the recommended Adenoma Detection Rate (ADR).</p>Yogesh ChaudhariAshish JaniHarshal A. SanghviAbhijit S. PandyaVipin GuptaDarshee Baxi
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s1427AI for 5G networking- A Bibliometric Analysis
https://ijisae.org/index.php/IJISAE/article/view/4834
<p>The merging of 5G networking with artificial intelligence (AI) has become a game-changing paradigm, completely overwhelming the landscape of communication systems. A thorough bibliometric analysis is presented in this work to help comprehend the developments, patterns, and significant research areas at the nexus of artificial intelligence (AI) and fifth-generation(5G). Through an extensive assessment of conference papers, patents, and academic literature, this investigation delves into the complex interactions that arise between artificial intelligence and 5G technology. This bibliometric analysis is a valuable tool for scholars, policymakers, and industry experts trying to understand the intricacies of AI-enabled 5G networks, as it offers a broad perspective of the scholarly scene. In this dynamic and quickly developing sector, it provides insights into the current level of knowledge, points out research gaps, and lays-out a plan for future studies.</p>Smita MahajanShivali Amit WagleSunil M. SangveAryani Gangadhara
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s2839Breast Cancer Detection Using Transfer Learning
https://ijisae.org/index.php/IJISAE/article/view/4835
<p>By illuminating the complex interactions between societal stigmas and gender inequities that frequently obstruct early diagnosis, this research constitutes a critical first step towards resolving the difficulties associated with breast cancer detection. Through a comprehensive analysis of risk factors, which encompasses the subtle impacts of biochemical pathways and underlying pathology, the study leverages the abundance of data found in datasets like CBIS-DDSM, SEER, and BreakHis to provide invaluable insights into breast cancer imaging. The research uses deep learning approaches, notably the MobileNetV2 architecture with transfer learning, and is a pioneer in the integration of cutting-edge technology. The results provide a respectable degree of precision in differentiating between benign and malignant cases, even with the intrinsic complexity shown in the 0.616 total accuracy score. A balanced f1-score and notable precision strengths for benign situations highlight the model's potential use in clinical settings. By highlighting the revolutionary potential of deep learning in improving diagnostic tools and changing the landscape of breast cancer detection, this research lays a solid platform for future developments.</p>Jyoti Pandurang KshirsagarBhagwan PhulpagarPramod Patil
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s4054MFAAMDTL: An Efficient Multimodal Feature Analysis Model to Mitigation Cloud Attacks using Transfer Learning Operations
https://ijisae.org/index.php/IJISAE/article/view/4836
<p>The persistent issue is that cloud applications fortified post-deployment with security patches remain susceptible to sophisticated attack vectors. In response to this, the discourse introduces an innovative, lightweight header layer designed to preemptively filter incoming requests prior to their processing by Cloud Virtual Machines (CVMs). Leveraging a combination of instantaneous and temporal analytics, this layer is adept at the early detection and neutralization of a broad spectrum of both active and passive cybersecurity threats, significantly bolstering the resilience of cloud deployments against malicious endeavors. To operationalize this defense mechanism, the system deploys an advanced logging framework capable of high-velocity data capture, triggered by an array of header-level events such as authentication attempts, access requests, and the temporal intervals between successive requests. This granular data collection strategy equips the system with a comprehensive dataset, derived from continuous user interactions, which is subsequently subjected to an intricate post-processing regimen aimed at the extraction of multimodal features. This process involves the manual tagging of request-response pairs by a curated group of users, facilitating the identification of diverse threat signatures such as temporal attack probabilities, IP-based attack typologies, user access patterns, and anomalies in request-response dynamics. At the heart of this model lies a sophisticated deep transfer learning framework, integrating the nuanced capabilities of Long Short-Term Memory (LSTM) networks and Gated Recurrent Unit (GRU)-based Recurrent Neural Networks (RNNs), trained on an extensive corpus of user-generated data. This hybrid RNN methodology enables the model to discern and classify a wide array of attack vectors with remarkable accuracy. An incremental learning module further refines the model's efficacy, enabling dynamic adaptation and continuous improvement in its predictive accuracy, precision, and recall metrics across various attack scenarios, including but not limited to Distributed Denial of Service (DDoS), brute force, cross-site scripting, SQL injection, as well as more passive threats like access control breaches and restricted ownership transfer attempts. Empirical evaluations of this model underscore its superior performance, achieving notable accuracy rates in detecting authentication attacks (99.3%), unauthorized access attempts (97.1%), DDoS and similar request-pattern aberrations (99.1%), and Man in the Middle (MITM) attacks (99.2%). When benchmarked against contemporary models, this innovative approach demonstrated a performance uplift of 6.5%, underscoring its viability for real-time deployment and scalability across diverse cloud networking scenarios.</p>Anagha RaichVijay Gadicha
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s5566Improved Deep Learning and Feature Fusion Techniques for Chronic Heart Failure
https://ijisae.org/index.php/IJISAE/article/view/4837
<p>Early detection of heart problems is of paramount importance, given that chronic heart failure remains a leading cause of global mortality. Accurate forecasting of cardiac conditions is crucial for timely intervention and improved patient outcomes. While various machine learning (ML) and deep learning (DL) models have emerged for cardiac disease diagnosis, most struggle to effectively handle high-dimensional healthcare datasets and often fail to significantly enhance chronic heart failure (CHF) diagnosis performance. In this study, we propose a smart healthcare framework that integrates deep learning and feature fusion techniques to predict CHF. Leveraging the PhysioNet datasets, our approach amalgamates features extracted from phonocardiogram (PCG) data. The study introduces novel algorithms, including lightweight CNN, hybrid CNN-autoencoder, and parallel hybrid CNN-autoencoder, offering promising avenues for enhancing CHF detection accuracy and efficiency. The performance of our proposed system is rigorously evaluated against alternative approaches, including feature extraction, machine learning, and traditional deep learning classifiers, using heart sound data. This research aims to advance CHF prediction capabilities, bridging the gap between cutting-edge technology and early cardiac healthcare intervention.</p>Namrata GawandeDinesh GoyalKriti Sankhla
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s6780Design of Efficient Pipeline Framework for Xml-Based Classifier Using Data Engineering Techniques
https://ijisae.org/index.php/IJISAE/article/view/4838
<p>In the field of modern era designing efficient framework for real time face detection systems stand out as innovative and technologically advanced solutions. This article describes the development and implementation of a system that leverages face detection model to accurately and efficiently identify objects in a variety of environments, including educational institutions, corporate environments, criminal detection and events. The xml-based face detection framework uses state-of-the-art using normal classification learning algorithms to analyze and recognize facial features, ensuring a high level of person identification accuracy. The framework can be seamlessly integrated into existing infrastructure, enabling an optimized and discreet monitoring of the image recording process. Additionally, the system is designed with user privacy and data security in mind, incorporating encryption and robust authentication mechanisms. Key features of the face detection system include real-time object detection, automatic data recording, and comprehensive reporting. The system’s user-friendly interface allows for easy integration into various organizational structures, making it a versatile solution for time and for checking the identity of various objects.</p>Nilesh D. NavghareL. Mary Gladence
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s8187Smart Marketing Investments: A Framework for AI-Based Financial Decision Support
https://ijisae.org/index.php/IJISAE/article/view/4839
<p>The use of artificial intelligence (AI) has become an important part of smart marketing investments in the fast-paced world of business and banking. Using AI's critical skills to improve accuracy, lower risks, and make the best use of resources is a huge step forward in this framework. As the amount and variety of data grows, standard ways of making decisions often don't work. We need a new way of thinking that uses AI to get useful insights.AI is very important for helping people make financial decisions because it can predict things better than humans can. AI quickly looks at huge information to find patterns, predicts market trends, and gives businesses an edge when making decisions. Being able to predict the future not only helps with strategic planning, but it also helps lower risks. AI models look at both past data and real-time market signs using machine learning algorithms. This lets businesses plan ahead for volatile market conditions and improve their financial stability. The proposed framework necessity of improving marketing funds in a time when allocating resources wisely is very important. With the help of data-driven AI models,and machine learning method for marketing budgets are carefully directed toward outlets and projects that are most likely to bring back the most money. This detailed method makes things run more smoothly, so businesses can quickly adjust to changing market conditions and get the most out of their marketing campaigns. This paper study about the bigger effects of AI-based financial decision support in the digital age, focusing on how it encourages new ideas, flexibility, and adaptation with morden machine learning methods. As companies try to figure out how to operate in today's complicated markets, this approach is a complete way to use AI's changing power to make smart and useful marketing investments.</p>Mahesh SinghManoj Kumar RaoSurendra S. JogiManoj B. PandeyAjit SaoMahesh Chopde
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s88100DRECQ: Design of an Efficient Model for Enhanced Diabetic Retinopathy Diagnosis Using Ensemble Classifiers and Deep Q Learning Process
https://ijisae.org/index.php/IJISAE/article/view/4840
<p>In addressing the critical need for advanced diagnostic tools in the realm of ophthalmology, particularly for the detection of diabetic retinopathy, this paper introduces a novel, ensemble-based approach, fusing the strengths of three distinct classifiers: Naive Bayes, Support Vector Machine (SVM), and Multi-Layer Perceptron (MLP). Traditional methods in retinal image analysis often fall short due to their static nature and inability to adapt to the unique complexities presented by individual images. This limitation manifests in less precise and accurate diagnostic outcomes, underscoring the urgent need for more dynamic and responsive techniques. The proposed model marks a significant departure from conventional approaches. By employing an ensemble method, it leverages the unique strengths of each classifier: the probabilistic analysis of Naive Bayes, the non-linear pattern recognition capability of SVM, and the intricate feature extraction proficiency of MLP process. The integration of these methods addresses the inherent limitations of using a singular approach, ensuring a more comprehensive analysis of retinal images & samples. Central to this innovation is the application of Deep Q Learning (DQL) for dynamic classifier selection. This reinforcement learning technique optimizes the ensemble by adaptively selecting the most suitable classifier for each specific retinal image, based on learned Q Values for different scenarios. This method not only enhances the accuracy and precision of diagnosis but also ensures continual adaptation and learning, keeping pace with evolving data patterns and advancements in imaging technology. The efficacy of this model is demonstrated through rigorous testing on the IDRiD & EyePACS Dataset. Results indicate a notable improvement over existing methods, with a 4.5% increase in precision, 5.5% in accuracy, 3.9% in recall, 4.9% in AUC (Area Under the Curve), 3.4% in specificity, and an 8.5% reduction in delay. These enhancements have profound implications for the field of ophthalmology. They signify a leap forward in the accuracy and timeliness of diabetic retinopathy diagnosis, ultimately leading to improved patient outcomes and a reduction in the burden on healthcare systems. This work, therefore, not only presents a technical advancement but also a significant stride in patient care, paving the way for more effective management and treatment of retinal diseases.</p>Minakshee ChandankhedeAmol Zade
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s101116Exploring ResNet101, InceptionV3, and Xception for Modi Script Character Classification
https://ijisae.org/index.php/IJISAE/article/view/4841
<p>The "MODI lipi" script, which was used historically in Maharashtra, Western India, to record religious writings, and which was the official script used by the Maratha administration from the 17th century until the mid-1900s, has a rich cultural legacy. Even though the "Manuscript" is a valuable source of inspiration and knowledge from a bygone age, modern audiences are not as familiar with it. Recognizing its potential to inspire and educate the present generation, there is a need to develop a sophisticated recognition system for MODI within handwritten character recognition. Deep learning-based algorithms, such as ResNet101, InceptionV3, and Xception, have demonstrated remarkable efficacy in various pattern identification applications, including character recognition. In the current landscape, transfer learning algorithms, especially those leveraging ResNet101, InceptionV3, and Xception architectures, have gained prominence for significantly enhancing recognition task outcomes. This study specifically proposes implementing these advanced deep-learning models to classify MODI characters. By harnessing the power of ResNet101, InceptionV3, and Xception algorithms, the aim is to optimize the recognition accuracy and efficiency of the MODI script, making it more accessible and comprehensible for today's youth. This research endeavors to unlock the untapped potential of MODI script as a valuable cultural and educational resource by utilizing state-of-the-art deep learning methodologies.</p>Ravindra SonavanePandharinath GhongeSandip Umajirao PatilKaustubh Shivaji SagaleAnand Arvind Maha
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s117124Statistical Investigation of Student Behaviour Analysis Models from An Empirical Perspective
https://ijisae.org/index.php/IJISAE/article/view/4842
<p>Student behaviour analysis is a multidisciplinary field which requires exploration of a wide variety of data, including, student’s geographical profile, area of behavioural study, temporal responses, situational responses, analytical reasoning, attention profile, etc. Combination of these factors requires design of intelligent machine learning approaches, which work on temporal behavioural responses. For instance, to predict student’s inclination towards technical education, models utilize analytical questionnaire, and social media tools to capture student’s behaviour. This data is processed using various deep learning architectures to estimate student’s inclination probability towards technical education. A wide variety of architectures are proposed for this task, and these architectures vary in terms of performance metrics, area of application, geography of student, etc. This makes it uncertain for researchers to test, validate &select most optimum models for their application, which increases cost & time needed for deployment. In order to reduce the uncertainty of model selection, this paper reviews some of the recently proposed methods for student behaviour analysis, and compares them in terms of performance metrics, area of application, and geographical parameters. The performance metrics include accuracy of analysis, computational complexity, mean squared error (MSE), and speed of analysis. This review will be helpful for researchers & behavioural analysis system designers to select the most optimum models for newer deployments, and will assist in performance upgradation of existing systems. Moreover, this text also recommends various improvements & enhancements in the reviewed models, which assists in upgrading their internal capabilities including scalability, flexibility, and performance analysis.</p>Ashwini RaipureSarika Khandelwal
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s125136XGBoost Learning for Detection and Forecasting of Chronic Kidney Disease (CKD)
https://ijisae.org/index.php/IJISAE/article/view/4843
<p>It's astounding that 63,538 cases have been documented. based on data from India's chronic kidney disease (CKD). Nephropathy in humans usually appears between the ages of 48 and 70. Compared to women, men are more likely to develop CKD. Regretfully, India has slipped into the top 17 countries for chronic kidney disease (CKD) since 2015. CKD is characterized by a gradual deterioration in the function of the excretory organs. Effective treatment and early illness identification may help prevent this terrible condition. Among other practical applications, machine learning is being used in fraud detection and medical research findings analysis. Chronic illness forecasting is done using a variety of machine-learning techniques. With a focus on decision trees, Adaboost, XGboost, random forests, logistic regression, support vector machines, naïve Bayes, KNN, and artificial neural networks, our primary goal is to evaluate the accuracy of various machine learning techniques. Here XGBoost ML algorithm performs well for prediction of chronic kidney disease (CKD), it provide the 99% accuracy and which is almost greater than the other ML algorithms tested. RCode has received praise from the study's performance analysis. This project's primary goal is to develop an application for sickness prediction that uses an analysis of the chronic kidney disease dataset to detect cases of chronic kidney disease (CKD) and non-CKD.</p>Yogesh KaleShubhangi RathkanthiwarP. FulzeleN. J. Bankar
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s137150Detection of ECG Wave Components for the Prediction of Acute Coronary Syndrome - Brief Survey
https://ijisae.org/index.php/IJISAE/article/view/4844
<p>An ACS (Acute Coronary Syndrome) is a term used to define the heart diseases like Heart attacks, Myocardial infarction, and Unstable Angina. The study described the Electrocardiogram is an important tool for measuring human health and disease detection. Electrocardiogram (ECG) signal consist of Components like waves, intervals and segments studied on the basis of time duration and size. PAN and TOMPKINS give the concept of QRS detection in the decade of eighty. Further several researchers developed various algorithms to detect QRS on the basis of derivative, wavelet transforms and other techniques. In this research we survey the progressive methods of detection of electrocardiogram wave components for the prediction of acute coronary syndrome by introducing electrocardiogram signal preprocessing, heartbeat segmentation, feature extraction and learning algorithms used. Additionally we depict some databases which is used for evaluation indicated by The AAMI standards were introduced by AAMI and are described in American National Standard Institute (ANSI/AAMI EC57:1998/(R) 2008) [16] for analyzing and describing the performance effect of cardiac rhythm and ST-segment evaluation algorithms. Sometimes monitoring and analyzing heartbeat ECG records are necessary. Most of the time there is a possibility of inaccuracy in ECG record analysis. This research becomes the alternative. It can provide essential information to doctors to carry out their diagnoses on patients.</p>Seema Mangesh ShendePrabhat Chandra ShrivastavaShrikant P. ChavateRatnesh RanjanSwati P. Aswale
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s151162Predicting Intraday Trend Reversals in Index Derivatives Using Supervised Machine Learning Algorithms
https://ijisae.org/index.php/IJISAE/article/view/4845
<p>This research paper delves into the realm of financial market forecasting, specifically focusing on predicting intraday trend reversals in index derivatives using supervised machine learning algorithms. The study encompasses a comprehensive examination of various machine learning techniques, including Support Vector Machines, Random Forests, XGBoost, and LSTM, to develop models capable of navigating the complexities inherent in the financial markets.The primary objective of the research is to enhance the predictive accuracy of stock market movements by incorporating a range of factors such as market conditions, liquidity, and external influences. This multifaceted approach aims to capture the dynamic and often unpredictable nature of financial markets, offering a more nuanced and effective prediction model.Through meticulous analysis and evaluation, the paper demonstrates the significant potential of machine learning technologies in the field of computational finance. It explores the strengths and limitations of each algorithm, providing an in-depth understanding of their applicability in real-world market scenarios.Furthermore, the research identifies key areas for future exploration, emphasizing the need for a more detailed examination of macroeconomic and sociopolitical factors, as well as the utilization of high-frequency data, particularly in emerging markets. These insights pave the way for ongoing advancements in the application of machine learning for financial market analysis.Overall, this paper makes a notable contribution to the field of computational finance, offering valuable perspectives and tools for academics and practitioners alike. It lays the groundwork for further research that aims to refine and expand the use of machine learning in stock market prediction, ultimately leading to more robust and versatile forecasting models.</p>Payas DeshpandeSridhar SubramanianShivali Amit WaglePreksha Pareek
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s163176In-Vitro Detection of Tooth Decay Using Reduced Graphene Oxide (rGO) based Sensor
https://ijisae.org/index.php/IJISAE/article/view/4846
<p>The pH in the tooth is related to the tooth cavity, which mainly arises due to the interaction of the cariogenic bacteria with the carbohydrates present in the food. Thus, in this work, to detect the tooth cavity we have fabricated the pH sensor which comprises the interdigitated electrodes (IDEs) mounted on the electrodes on the printed circuit board. Fabricated IDEs have length and width of about 100 mm with 0.05 mm of copper acting as the electrodes. Further, rGO is used as the sensing film to detect the pH. For this purpose, rGO in powder form is converted into the liquid with the help of dispersion in the ethanol followed by sonication for about 20 minutes. Then the rGO sample is drop casted on the fabricated IDEs and subsequently air dried for 24 hrs. Further, we have taken the standard pH solution ranging from 4 to 12 and measured the sensor resistance. Further, to enhance the accuracy of the detection we have implemented the principle component analysis and k-mean algorithm on the self-collected data.</p>Payal GhutkeWani Patil
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s177183Secured Authentication Using ECC Based Fractal Fuzzy in Cloud
https://ijisae.org/index.php/IJISAE/article/view/4864
<p>In this paper, the study uses a novel design of the fractal fuzzy model in devising the secured authentication process via Elliptic Curve Cryptography (ECC). This model assists in improving the process of authentication in distributed cloud systems. The study uses attributed based data segregation that is developed using the preference of the data owners or users. Here, the priority is offered to sensitive data rather than non-sensitive data. After the grouping of the sensitive information based on the attribute, these subgroups are encrypted via group keys. Such encrypted information is then merged with the attributes that are of a non-sensitive type and then it is uploaded to the distributed cloud environment. The simulation is conducted in a cloudsim simulator to test the efficacy of the model against various types of attacks. The results of the simulation show that the proposed model performs with minimal computational complexity, storage space and processing time than the existing authentication models. Further, it offers an increased rate of security over sensitive and non-sensitive information than other methods.</p>A. Ahadha Parveen P. S. S. Akilashri
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s184194An Innovative Reliable Client-Centric Deep Learning Inference Methodology
https://ijisae.org/index.php/IJISAE/article/view/4865
<p>Mobile phones and tablets have access to a very huge amount data that may be utilized to train learning models, potentially improving the user experience significantly. Nevertheless, the data available is often both extensive and sensitive, making it challenging to collect at centralize server and train within a centralized server using conventional methods. In this study, we investigate the utilization of blockchain technology with decentralized digital ledger to create a decentralized client-centric distributed learning system with the flexibility to support various machine learning models. This system enables the training of machine learning models directly on local machines, thereby addressing the constraints imposed by centralized servers. We demonstrate our system design, which includes two decentralized blockchain models built using Python Tensor Flow to ensure the system's reliability and efficiency. Ultimately, Block-CCL serves as an experimental environment for evaluating and distinguishing the impact of decentralized client centric i.e. federated learning from synchronization of model methods on the performance of the entire system. This highlights the validity and effectiveness of a federated learning system as a viable alternative to more centralized machine learning models.</p>Ganesh D. GovindwarSheetal S. Dhande
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s195201E-Voting Based Blockchain Mechanism Using Feature Selection Based Machine Learning
https://ijisae.org/index.php/IJISAE/article/view/4866
<p>The issue of scalability in e-voting system is typically circumvented by the utilization at majority of the time. These components are capable of operating without needing one another in order to function, and they are even able to keep their own records independently of one another. In this paper, we develop an intelligent E-voting system that uses machine learning algorithm for feature selection to generate the QR codes. The aim is to improve the security of the QR based on features by optimal selection of features. The QR code with other registered details is stored in database via proper blockchain authentication mechanism. The simulation is conducted in python to test the efficacy of the model against various other methods and the results of simulation shows that the proposed method has a reduced computational complexity and increased accuracy on feature selection than other existing E-voting mechanisms.</p>T. PrabakarS. Kanchana
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s202212Design of IoT-Based Wearables for Health Care Prediction Using Normalized-Patch Gan Based Fruit Fly Optimization
https://ijisae.org/index.php/IJISAE/article/view/4867
<p>Healthcare is a critical sector where timely and accurate predictions can save lives and improve the quality of care. Traditional healthcare systems often lack the ability to process vast amounts of patient data efficiently. To address this, IoT technology is harnessed for seamless data collection and integration, facilitating real-time updates to a central database. The challenge lies in harnessing this data effectively to predict health conditions. The diversity of patient data, including Medical IDs, pulse rates, medical reports, and symptoms, requires sophisticated algorithms to extract meaningful insights. Moreover, the accuracy and reliability of predictions are vital to ensure patient safety. This paper presents the design of an Internet of Things (IoT)-based healthcare prediction system utilizing the Normalized Patch Generative Adversarial Network (NP-GAN) based Fruit Fly Optimization (FFO) algorithm. The proposed system aims to predict health conditions based on patient data, including Medical ID, pulse rate, medical reports, and symptoms. Through seamless integration of IoT technologies and AI algorithms, the system enables real-time monitoring and predictive analysis, enhancing patient care and medical decision-making. The system collects patient data including Medical ID, pulse rate measurements, medical reports, and reported symptoms. IoT devices facilitate real-time data transmission to the central database. Raw data undergoes preprocessing, including normalization and sequence alignment. Textual medical reports are transformed into numerical vectors using techniques like word embeddings. Features such as pulse rate trends, symptom sequences, and medical report patterns are extracted from the preprocessed data, providing valuable insights for prediction using NP-GAN. The RCNN algorithm, combining recurrent and convolutional layers, is employed for its ability to capture temporal dependencies and spatial patterns in data. The network learns to associate pulse rate trends, symptoms, and medical information for accurate predictions. The RCNN model is trained using historical patient data and validated using FFO to optimize hyperparameters and prevent overfitting. Real-time patient data is continuously fed into the trained RCNN-FFO model, which predicts potential health issues. Alerts are generated for medical professionals if anomalies or concerning patterns are detected. The system performance is assessed using metrics like accuracy, precision, recall, and F1-score. Continuous feedback and retraining improve prediction accuracy over time.</p>T. S. BaskaranC. Veeraprakashkumar
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s213224A Hybrid Cluster Based Intelligent IDS with Deep Belief Network to Improve the Security over Wireless Sensor Network
https://ijisae.org/index.php/IJISAE/article/view/4868
<p>Numerous inexpensive, compact devices compose a Wireless Sensor Network (WSN). They're usually readily available to some types of attacks due to their location, which is not well protected. A large number of researchers are focusing on WSN security at the moment. This kind of network is characterized by vulnerable characteristics, such as the ability to organize oneself without a stable infrastructure and open-air transmission. To train variables for the probability-based feature vectors, a Deep Neural Network (DNN) framework that is derived from international vehicle network packets shall be applied. The detector is capable of detecting any malicious attack on the vehicle since DNN gives each category a chance to distinguish between attacks and regular packets. Intrusion Detection Systems (IDS), can help to identify and stop security attacks on vehicles. The study proposes a mechanism for enhancing the security of WSNs based on Hybrid Clusters and Intelligent Intrusion Detection Systems with Deep Belief Networks (HCIIDS-DBN). It can provide a protection system for intrusions and an analysis of vehicle attacks in real time. They are designed based on their respective attack probability and ability, to the sensor node, sink, or cluster head. The proposed HCIIDS-DBN is composed of modules designed to detect anomalies and dereliction. The objective is to increase detection rates and decrease false positive incidences by detecting anomalies and abuse. Finally, the detected data are integrated and the various types of vehicle communication attacks are reported using the Decision Support System (DSS). The results of the experiment show that the proposed method may respond to the attack in real-time with a much detection of higher ratio in the Controller Area Network (CAN) bus.</p>Priyanka R.Teena K. B.Rashmi T. V.Reshma J.Tejashwini NagarajTejaswini N.
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s225238Optimizing Cryptocurrency Price Prediction: A Hybrid Approach with Resilient Stochastic Clustering and Gravitational Search Algorithm
https://ijisae.org/index.php/IJISAE/article/view/4869
<p>Cryptocurrency markets have become increasingly complex, making accurate price prediction a challenging task. This article proposes a Hybrid Oppositional Sparrow Search of Gravitational Search Algorithm (HOSS-GSA) which separated iterative chaotic routing to address problems of its probability of falling optimal solutions. The proposed hybrid framework aims to harness the strengths of each component to improve prediction accuracy and capture the dynamics of cryptocurrency historical price data. Resilient Stochastic Clustering effectively identifies relevant features and reduces dimensionality, enhancing the efficiency of subsequent prediction steps. Furthermore, it helps in identifying clusters of similar data patterns within the cryptocurrency historical prices dataset. HOSS-GSA aims to optimize model parameters and improve the overall performance of the prediction model. The experiments were conducted by the common evaluation operations to validate the functionality of grouping for high-dimensional multiview information of Cryptocurrency Historical Prices Dataset, as well as the Wilcoxon rank-sum assessment model was used to measure variable influence for the technique, outperforms traditional prediction methods, achieving higher prediction accuracy and robustness. This approach provides a valuable tool for cryptocurrency traders, investors, and analysts seeking to make informed decisions in a rapidly evolving market.</p>R. RameshM. Jeya Karthic
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s239248Fuzzy Integrated Latent Dirichlet Allocation Algorithm for Intrusion Detection in Cloud Environments
https://ijisae.org/index.php/IJISAE/article/view/4870
<p>This research presents an in-depth exploration of the FI-LDA model, showcasing its efficacy in anticipating and preventing intrusions, thereby bolstering security measures within cloud environments. The study introduces a novel approach to intrusion prevention, fostering a robust predictive model that significantly enhances the system's capability to discern evolving attack patterns. Leveraging fuzzy modeling, the research demonstrates the utilization of vast amounts of unlabeled data, resulting in heightened accuracy and reliability of the system. The evaluation of diverse elements crucial for cybersecurity underscores the comprehensive approach adopted to achieve the research objectives. While the FI-LDA model exhibited a favorable trade-off, addressing a pervasive flaw, there remains a call for further refinement to detect assault patterns more effectively. The research concludes by highlighting the commendable effectiveness of the FI-LDA model in identifying and detecting malicious activities within the cloud environment, affirming its strong overall performance and contribution to advancing intrusion detection systems.</p>Vinayak Kishan NirmaleC. Madhusudhana RaoMylapalli RameshM. NirmalaS. GirinathNellore Manoj Kumar
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s249259Optimizing trust, Cloud Environments Fuzzy Neural Network, Intrusion Detection System
https://ijisae.org/index.php/IJISAE/article/view/4871
<p>In the dynamic landscape of cloud computing, ensuring the security and integrity of services is paramount. This article introduces a novel approach to cloud intrusion detection by leveraging the synergies of fuzzy logic and neural networks. The proposed Fuzzy Neural Network Aided Cloud Intrusion Detection System (FNN-CIDS) integrates the adaptability of fuzzy systems with the learning capabilities of neural networks to enhance the detection accuracy of malicious activities within cloud environments. The system is designed to discern subtle patterns indicative of intrusion attempts, thereby fortifying the defense mechanisms for trusted services hosted in the cloud. The article presents the conceptual framework of FNN-CIDS, detailing the integration of fuzzy logic for rule-based inference and neural networks for pattern recognition. Experimental results demonstrate the system's efficacy in identifying diverse intrusion scenarios while minimizing false positives. This research provides a promising path for improving the reliability of cloud computing infrastructures and advances strong security frameworks for cloud-based applications. In this sense, the research effort provides a trust evaluation system to determine the reliability of cloud services and an intrusion detection system to guarantee intrusion-free cloud services. the construction of a cloud intrusion detection system using a neuro-fuzzy based self-constructing clustering algorithm. The performance of this method has been compared to other well-known clustering methods in the field of cloud intrusion detection using result analysis.</p>Archana B.N. JeebaratnamB. Nageswara RaoU. SesadriN. ShirishaNellore Manoj Kumar
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s260275Comparing the Accuracy of CNN Model with Inception V3 for Music Instrument Recognition
https://ijisae.org/index.php/IJISAE/article/view/4872
<p>Identification of music instruments from an audio signal is a complex but useful task in music information retrieval. Deep Learning and traditional machine learning models are extremely very useful in many music related tasks such as music genre classification, recognizing music similarity, identifying the singer etc. Music Instrument recognition and classification would be helpful in categorizing different categories of music. Many researchers have proposed models for classifying western music instruments. But very little research has been done in identifying instruments accompanied with South Indian music. This research aims at identifying string instrument such as violin and woodwind instrument such as flute accompanied in a Carnatic music concert and also in other categories of music. In order to identify the instruments accompanied, Convolutional Neural Network model and Inception V3 models were used. The Mel Frequency Cepstral Coefficients images were extracted from the audio input and fed in to the neural network model. The model has been trained for the above mentioned instruments, tested and validated on different types of audio input. This research also evaluates the performance of Inception V3 transfer learning model with CNN model in recognizing the instruments used in different categories of music.</p>Renju K.Ashok Immanuel V.
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s276282IoT Based Agriculture Monitoring and Prediction of Paddy Growth using Enhanced Conquer Based Transitive Clustering
https://ijisae.org/index.php/IJISAE/article/view/4874
<p>Agriculture is widely recognized as a fundamental pillar of our civilization, and it is currently undergoing a significant transition with the emergence of the IoT. This research investigates the field of IoT-based agriculture monitoring, with a specific emphasis on forecasting paddy growth. The introduction establishes the context by emphasizing the pivotal significance of agriculture and the promise of the IoT in enhancing farming methodologies. The problem statement highlights the necessity for a more advanced and precise system to monitor and forecast the growth of paddy, by identifying a gap in current research. Conventional approaches frequently prove inadequate in delivering timely and comprehensive insights, so neglecting to fully exploit the capabilities of IoT technology. The Enhanced Conquer based Transitive Clustering methodology combines conquer-based methodologies with transitive clustering, providing a resilient framework for the study and prediction of data. By harnessing the capabilities of IoT devices, real-time data pertaining to many parameters, including soil moisture, temperature, and humidity, is gathered. The study findings demonstrate the effectiveness of the Enhanced Conquer based Transitive Clustering algorithm in properly forecasting paddy growth stages. The system possesses the capability to not only monitor the prevailing agricultural circumstances but also forecast forthcoming developments, thereby empowering farmers to make well-informed decisions. The model accuracy and effectiveness highlight its potential for extensive implementation in contemporary agricultural practices.</p>C. MuruganandamV. Maniraj
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s283293Cloud Dynamic and Public Auditing Scheme for Secure Data by using RSA with Modified Dynamic Hash Table
https://ijisae.org/index.php/IJISAE/article/view/4875
<p>Users are able to host and access a broad variety of internet-based services thanks to a paradigm for distributed computing that is known as "cloud computing." It sees widespread deployment in corporate applications like as data storage and internet-based software programmes. Users of cloud storage services are able to access their personal information whenever they want and from any place they want since they are not required to keep the data locally. The cloud service provider, on the other hand, does not put their whole faith in cloud storage since the integrity of the data is reviewed on unstable cloud servers. A significant amount of research has been conducted in public auditing with the goal of lowering the amount of computation time required for the integrity check. The method that is used the most has low costs associated with the processing but offers no security. In the work that we have suggested, a component known as the Cloud Auditor (CA) is responsible for carrying out the task of monitoring any changes that may have been made to the block by an external hacker; original data is obtained from its cache record. The TPA is where the data information required for dynamic auditing is stored, and it is recorded in the TPA's modified dynamic hash table (MDHT). The information is encrypted via the MRSA algorithm, which is a modified version of the RSA algorithm. The MDHT is not the same as a dynamic hash table since the former does not include a tag block while the later does. The cost of calculating the MDHT is analysed, then compared to other approaches that are currently being used. The results of the experiments showed that the MDHT method has a reduced computation cost compared to the most current public auditing methodologies.</p>A. Ahadha ParveenP. S. S. Akilashri
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s294303Oil Spill Detection and Recognition Utilizing Faster R-CNN with Enhanced Mobilenetv2 Architecture
https://ijisae.org/index.php/IJISAE/article/view/4876
<p>Oil spills are a major hazard to the environment, animals, and ecosystems. The identification of oil spills in a timely and accurate manner is critical for successful mitigation and response operations. We offer a complete strategy to oil spill detection and identification in this study by incorporating modern computer vision techniques. First, to improve picture quality and minimize noise in the input data, we use a Non-Adaptive Threshold with Contrast Limited Adaptive Histogram Equalization (CLAHE). This preprocessing procedure increases overall picture quality, making subsequent analysis more trustworthy. Then, using a Fused UNet Segmentation model, we apply the power of deep learning to picture segmentation. This approach efficiently isolates oil spill sites from the backdrop, allowing for exact identification and study of polluted areas. We use a Convolutional Neural Network (CNN) based on the well-known AlexNet architecture to extract relevant features from segmented photos. This stage extracts discriminative features, which improves the model's capacity to differentiate between oil spill and non-spill locations. The combination of Faster R-CNN with Enhanced MobileNetV2 architecture is at the core of our suggested solution. This hybrid approach delivers not only real-time processing but also cutting-edge performance in object identification tasks. We allow our model to identify and characterize oil spills correctly and effectively by training it on a dataset that includes both synthetic and real-world oil spill photos. To deliver a comprehensive solution for oil spill detection and identification, we integrate cutting-edge picture enhancement, segmentation, feature extraction, and object detection approaches. Experiment findings show that the system is excellent at detecting oil spills in a variety of environmental situations, allowing for faster reaction and mitigation measures to safeguard our valuable ecosystems. By employing advanced computer vision techniques, our system aligns with SDG 6 (Clean Water and Sanitation) by safeguarding water resources through accurate detection of oil spills. With a focus on SDG 14 (Life below Water), our technology aids in the preservation of marine ecosystems by minimizing the impact of oil spills on aquatic life.</p>V. SudhaAnna Saro Vijendran
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s304321Brain Computer Interface Based Home Automation System
https://ijisae.org/index.php/IJISAE/article/view/4877
<p>Home automation is being developed with the aid of Internet of Things (IoT) technology to control any devices by connecting them to the internet. All home appliances are able to operated with the help of IoT technology. But for someone with physical limitations, using the appliances without assistance is still a difficult undertaking. The BCI (brain controlled interface) technique enables people with physical disabilities to interact with household equipment. A brain and a machine can work together to create a form of brain-computer interface (BCI) that allows signals through the brain to operate an outside action, like moving a cursor or adjusting a prosthetic limb. It is possible to control an object directly from the brain thanks to the interface. The technology will translate the brain's messages into actions by interpreting them. With the aid of a Bluetooth module, these electrical signals are transformed into their original waveform and handled in the direction of an Arduino controller. The chip in the Arduino will continue to process the data that was received before utilizing a relay to control the appliances. The system that is suggested will be easier to build and more cost-effective.</p>M. MeyyammaiP. Marish KumarK. A. Indu Sailaja
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s322328Sign Language Recognition Using Convolutional Neural Network
https://ijisae.org/index.php/IJISAE/article/view/4878
<p>Communication, the sharing of information, ideas, and feelings, is typically facilitated through a common language. However, for individuals who are deaf and mute, communication presents unique challenges due to the inability to hear or speak. Sign language emerges as a crucial medium for communication among the deaf and mute and with those who can hear and speak. Unfortunately, the broader population often underappreciates the significance of sign language, resulting in communication barriers.</p> <p>To address this communication gap, we propose a machine learning solution—an innovative model designed to recognize various sign language gestures and translate them into English. Current Indian Sign Language Recognition systems, while employing machine learning algorithms, often lack real-time capabilities. In this paper, we introduce a method to construct an Indian Sign Language dataset using a webcam. We then leverage transfer learning to train a TensorFlow model, culminating in the development of a real-time Sign Language Recognition system. Notably, this system demonstrates commendable accuracy, even with a relatively modest dataset.</p> <p>The creation of a real-time sign language detector marks a significant stride in improving communication between the deaf and the general population. We proudly present the implementation of a sign language recognition model, rooted in a Convolutional Neural Network (CNN) and utilizing a Pre-Trained SSD Mobile-Net V2 architecture. Applying transfer learning, we have achieved a robust model consistently classifying sign language gestures. This groundbreaking innovation not only facilitates effective communication but also serves as a valuable tool for individuals learning sign language, providing practical opportunities for practice.</p>Gayathri D.Azhagu RamarS. KarpagamSudharsanan J.Rishwan S.Sudalaimani P.
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s329337A New Correction Factor-based Strategy with DWT-SVD for Contrast Enhancement in Digital Mammograms
https://ijisae.org/index.php/IJISAE/article/view/4880
<p>Globally, breast cancer stands as the second most prevalent disease affecting women. Mammography, utilizing low-dose X-rays, remains a highly effective modality for the early detection of cancer. Challenges such as uneven illumination and machine-imposed limitations contribute to low-contrast mammogram images, potentially impacting the accuracy of diagnoses. Due to the inherently narrow intensity range in mammography images, distinguishing between cancerous and non-cancerous tissues becomes challenging. This paper introduces a novel approach that combines Adaptive Gamma Correction with a two-way Discrete Wavelet Transform-Singular Value Decomposition (DWT-SVD) to enhance the visual clarity of the resulting images while preserving crucial clinical information. The introduction of a new correction adjustment factor enhances the singular value of the image, resulting in a significantly improved contrast-enhanced output. Experimental validation is conducted using mini-MIAS dataset, assessing the proposed technique with quantitative parameters such as Structural Similarity Index Measurement (SSIM), Pearson Correlation Coefficient (PCC), Peak to Signal Noise Ratio (PSNR), Contrast Improvement Index (CII), Mean Absolute Error (MAE), and Average Mean Brightness Error (AMBE). The obtained average values, including scores of 0.929, 0.998, 22.875, 1.136, 14.457, and 14.138, respectively, demonstrate promising results compared to conventional methods. Furthermore, comparison with the state-of-the-art techniques shows improved results, showcasing significant advancements in local information preservation and contrast enhancement in mammography images.</p>Dharmendra KumarAnil Kumar SolankiAnil Kumar Ahlawat
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s338352Innovative Graph Convolutional Neural Networks for Probing Aphasic Functional Connectivity in fMRI Data
https://ijisae.org/index.php/IJISAE/article/view/4881
<p>In neuroscience, exploring the role of brain connectivity in language processing was fundamentally important. Recent developments in feature extraction, the insights offered by transformer-based language models, and comprehensive approaches to studying acute ischemic stroke underscored the urgency for groundbreaking research methods. Addressing this need, our study introduced the Innovative Graph Convolutional Neural Network (GCNN) Paradigm. This novel approach was adept at examining aphasic functional connectivity, utilizing the capabilities of advanced Functional Magnetic Resonance Imaging (fMRI) data analysis. This research adopted an all-encompassing strategy. It leveraged a varied group of participants and state-of-the-art imaging technology, notably the Siemens Prisma 3 Tesla MRI scanner. Our methodology was meticulous, involving detailed data collection, a comprehensive preprocessing routine, and the deployment of our groundbreaking GCNN framework. We adhered to a training, validation, and testing division of 70-15-15%. The evaluation of the model was thorough, employing metrics like accuracy, precision, recall, and F1 score, and was further strengthened by a 5-fold cross-validation approach. Our findings indicated significant changes in brain connectivity associated with aphasia. The GCNN model excelled in both performance and clinical relevance, marking a substantial step forward in our understanding of how neural networks facilitate language processing. The precision of the GCNN Paradigm not only enhanced our grasp of these neural networks but also set new precedents for meticulousness and ethical standards in scientific research.</p>Kailash Nath TripathiSudhakar TripathiRavi Bhushan Mishra
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s353362Location Aware Content Priority based Recommendation System Flying Squirrel Optimization - Deep Alternative Neural Network (FSO-DANN)
https://ijisae.org/index.php/IJISAE/article/view/4882
<p>Social networks collect a lot of customer information, and this data may be utilized to develop knowledge for a variety of mobile and web applications. The Recommendation System (RS) is a domain garnering much attention these days. There are numerous itinerary and location aware content in RS available right now, some of that are nearly exclusively business. However, a thorough analysis demonstrates the need for study and advancement in this field. The data supporting this study show that the majority of systems are essentially destination RSs, and the great majority do not dynamically build routes but instead require the customer to choose appropriate locations. Some need greater user involvement while others fail to account for the length of presence at the selected sites. In certain frameworks, the community finding technique was ineffective, while in others, the routes are not the best. The RS was developed to fill the holes in the existing itinerary RS that were discovered through a thorough analysis. A Flying Squirrel Optimization - Deep Alternative Neural Network (FSO-DANN) based location RS was in charge of giving customers to get the priority of the best location. A backtracking-based approach is implemented in the genetic algorithm-based itinerary planning component to create itineraries. The Hadoop Map Reduce programming method was used to build the system in parallel. An extensive investigation of the system's assessment reveals that it is effective and competent enough to offer a trip schedule to a user that was better suitable in 25% less time than existing systems.</p>P. Christopher ArokiarajD. Hari Prasad
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s363369A Novel Deep Learning Approach for Greenhouse Crop Growth Prediction
https://ijisae.org/index.php/IJISAE/article/view/4883
<p>The precise management of environmental conditions ensures increased crop production, and crop growth prediction in greenhouses plays a big part in agricultural design and governance in greenhouses. Using growth prediction in greenhouses, growers and farmers can better plan for the future and save money. But, it's a very tough process. Radiations, CO<sub>2</sub>, temperature, condition of seedlings, soil conditions and fertilization, illness rates, and many other aspects all affect crop production in a greenhouse. A wide range of factors affect crop output, and it's not easy to build a precise model that accounts for all of them. This investigation makes use of a novel Bayesian optimized artificial neural network (BOANN) to predict the development of greenhouse crops. For this study, diverse datasets of greenhouses from various periods are gathered and preprocessed using min-max normalization to standardize the raw data. Kernel-based principal component analysis (K-PCA) and the wrapper technique are used, respectively, for feature extraction and feature selection. The experimental outcomes of datasets gathered from greenhouses over a range of periods demonstrate that the proposed BOANN approach outperforms other existing approaches in terms of prediction rate, mean square error (MSE), f1-measure, and recall.</p>E. Mercy BeulahS. Radha RammohanSangeetha VaradhanV. VaidehiK. Anuradha
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s370380Design and Development of Data-Driven Product Recommender Model for E-Commerce using Behavioral Analytics
https://ijisae.org/index.php/IJISAE/article/view/4884
<p>Recommendations assist users in more precisely locating the information they require for a given sample. People all around the world have been drawn to E-Commerce-based businesses in recent years. The Recommendation Model (RM) is an important system in internet business that recommends products to consumers based on their previous actions. Furthermore, the RM is effectively employed by both corporate service suppliers and customers. Furthermore, because so much product information exists online, recommender systems are critical for analyzing the existence of items that should be offered to clients, which enhances customer decision-making by giving extensive knowledge about the product and saves the effort required. However, the complications are recognized and observed from various methodologies as per the literature. To maintain proper RM, the research needs to focus more on data collection and analysis that provide real-time support. Thus, the user behavior data and machine learning concepts are utilized for designing Data-Driven Product Recommender Model (DD-PRM). From the experimental results, it has been determined that the proposed DD-PRM outperforms than the exiting models.</p>D. Srinivasa KumarKilaru MadhaviTenneti RamprasadK. R. SekharSrinivasa Rao DhanikondaCH Ravi
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s381392Evolutionary Optimization of Dominating Set-Based Virtual Backbone Cluster Scheduling for Enhancing Energy Efficiency in Asymmetric Radio WSNs
https://ijisae.org/index.php/IJISAE/article/view/4886
<p>Wireless Sensor Networks (WSNs) are becoming essential for many uses, such as industrial automation and environmental monitoring. To extend the network's lifespan, energy efficiency is critical. Asymmetric radio WSNs pose a special difficulty in energy consumption optimization since nodes in these networks have different transmission capacities. This study presents a unique method for improving energy efficiency in asymmetric radio WSNs using Genetic Algorithm-based Dominating Set-Based Virtual Backbone Cluster Scheduling (GADS-VBCS). By dividing the network into clusters using the idea of virtual backbones, the proposed GADS-VBCS method efficiently lowers communication overhead. The programme considers the asymmetry in radio ranges and effectively plans the activation of clusters based on a dominating group of nodes using evolutionary algorithms. This scheduling ensures network coverage and connectivity while optimizing energy consumption. To assess GADS-VBCS's performance, comprehensive simulations across several situations are carried out and compared with current methodologies. The findings show that, especially in asymmetric radio WSNs, GADS-VBCS performs better than traditional scheduling methods regarding energy efficiency, network overhead, network lifetime, and packet delivery ratio. This work provides a useful tool for real-world deployments in resource-constrained contexts by solving energy efficiency issues in asymmetric radio WSNs.</p>Venkateswarlu MannepallyBanitamani MallikK. AnuradhaPratiksha G. PatilD. KavithaSrinivasa Rao Dhanikonda
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s393403Wearable Computing: Canonical Correlation Analysis (CFA) Statistical Method to Validate the Measurement Models Smart Ergonomic Shoes
https://ijisae.org/index.php/IJISAE/article/view/4889
<p>Wearable sensors have become increasingly popular over the past few years, and there are now a number of different products on the market that may be used for monitoring one's own health and activities. A more modern health care system ought to be able to provide better medical services to people at any time and any location, in a manner that is both inexpensive and patient-friendly. This study intends to evaluate the measuring model and offer evidence for the links between influential factors and purchase intention. To do so, it will make use of CFA. The results of the CFA will contribute to a better knowledge of the elements that have a substantial impact on purchase intention in the context of smart ergonomic footwear, since this understanding will be furthered by the findings of the CFA. In general, canonical correlation analysis (CFA) is an effective statistical method that contributes to the validation of measurement models, the evaluation of correlations between variables, and the improvement of the robustness of the study's conclusions. The coefficient for Customer Perceived Values is statistically significant at the 0.05 level (since the p-value is less than 0.05). The positive coefficient suggests that there is a positive relationship between Customer Perceived Values and the outcome variable (purchase intention) in the analysis. The coefficient for Personal Values is statistically significant at the 0.05 level. The positive coefficient suggests that there is a positive relationship between Personal Values and purchase intention in the analysis. The coefficient for Social Factors is highly statistically significant (p-value is close to 0) at the 0.05 level.</p>Rakhi NagpalPritpal SinghPrikshat Kumar AngraGagandeep Singh CheemaMuzzamil Rehman
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s404408A Review and Research Panorama on Food Recommender System Based on Health Care
https://ijisae.org/index.php/IJISAE/article/view/4890
<p>In recent years, food recommendation systems have garnered increasing attention due to their pivotal role in promoting healthy lifestyles<strong>. </strong>Much of the current research in the food industry is dedicated to devising strategies for recommending suitable food products based on user preferences, health considerations, or a combination of both. These systems offer users the ability to not only receive personalized food recommendations but also to monitor their nutritional intake, encouraging them to make constructive changes to their dietary habits. This paper aims to provide a comprehensive overview of various recommender systems in the domain of recipe recommendation. Furthermore, it conducts a systematic review of the diverse contributions made in the field of food and diet recommender systems, considering user preferences, health factors, or a fusion of both. Additionally, the paper delves into the research challenges faced, the datasets employed, and the methodologies applied in the development of these food recommender systems</p>Gamini DhimanGaurav GuptaBrahmaleen K. Sidhu
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s409422A Survey Integration of Wearable Device with Block Chain Technology in the Healthcare
https://ijisae.org/index.php/IJISAE/article/view/4891
<p>This survey aims to explore the synergies between wearable devices, EHRs, and block chain technology in the context of generating digital passports. By examining the current state of research, technological advancements, and real-world implementations, the survey seeks to provide a comprehensive understanding of the opportunities, challenges, and implications of this innovative intersection. Through this exploration, we aim to contribute insights that will propel the integration of wearable’s and block chain-secured EHR digital passports into mainstream healthcare, fostering a future where individuals have greater agency over their health data and healthcare providers are empowered with more comprehensive and timely information for improved patient care.</p>Sathishkumar MV. Raghavendran
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s423426A Susceptible Evidence Processing Framework for Handheld Devices Through Digital Forensic Measurements
https://ijisae.org/index.php/IJISAE/article/view/4894
<p>Digital forensic comprises various actions for processing digital evidences like preprocess, identification, modeling, extraction, and documentation. All these actions are modelled and entitled through the court of law. Different procedures and methods are followed to perform these actions by the help of various platforms and hardware specifications. The analysis and processing of digital evidences depends on the hardware specifications of various companies and the systematic approach of various effective evidence processing software tools. Most of the hardware developing companies takes the security measures through on board circuits and this helps the digital investigators an advantage while retrieving evidences. Latest technological advancements in industry demands various sensitive security measures needs to be considered while launching new hardware devices specifically for communication purposes. Digital forensic plays a great role in retrieving sensitive evidences and its processing while a digital crime scene is evaluating. This activity considers various processing steps and it leads to the evaluation of both hardware and software participated in the crime scene. Mobile devices are the most sensitive and popular handheld devices used around the globe and the communication capability of these handheld devices makes the message passing and content delivery more flexible hence may lead to the misuse and hacked through the personal space. This article gives an effective framework for analysis and processing of digital evidences specifically for handheld devices like Mobiles, pager, laptop, Notebook and other electronic pads. Nowadays most of the communications occurred through handheld devices so the application of digital forensic measurements on these cases are highly important and sensitive. The digital crime analysis and its effective processing solved by the proposed framework and it integrates various levels of security pads. The framework proposed here comprises LR based Numerical and Verbal likelihood ratio during the digital evidence processing scenarios. This integrated mechanism works on the device platform scrutinize both platform dependent and independent factors and applied on the kernel layer with certain security measurements. Any handheld or mobile platforms may adapt with the changes and the retrieved kernel resources including any suspected communications can pass through the framework channel. Thus the scalable platforms may arise with sustainable security enhancements which are entitles according to the procedure established by law.</p>T. M. BhraguramP. S. RajakumarN. Kanya
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s427442Employee Segmentation by Measuring the Attitude of ‘Intention to Stay’: A Machine Learning Approach
https://ijisae.org/index.php/IJISAE/article/view/4903
<p>This study aims to segment the employees based on their attitude towards more extended stay in the organization. Employees' attitude differs despite having an elite HR system and administration. However, identifying the factors that influence the employees' attitude towards their stay is crucial to retaining them. Hence, this study measures their attitudes regarding working conditions, supervision, compensation and benefits, task assignments, amenities, grievance handling system, and other HR operating factors. As a result, this study uncovers the position of employees into different segments based on this measurement task. The samples are drawn from the shop-floor and operation level of employees working in a textile company with more than 500 crores as annual turnover.</p>Ramkumar S.Karthika P.D. DilipMahesh Singh
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s443449A Novel Corona Graph Based Proof-of-Work Algorithm for Public Blockchains
https://ijisae.org/index.php/IJISAE/article/view/4904
<p>Various mathematical puzzle-based Proof-of-Work (PoW) algorithms have been developed to prevent adversaries from adding illegitimate nodes in public blockchains. The puzzles are devised in such a manner that they are difficult to solve by the blockchain members, hereto called ‘provers’, but easily verifiable for correctness by the blockchain administrator, hereto called as ‘verifier’. However, existing algorithms are either very computationally intensive requiring high end processing capability or consume a lot of space for storing homogeneous data tables at prover and verifier ends. In this article, we introduce a novel Corona Graph based PoW technique for distinguishing adversaries from legitimate blockchain clients and hence mitigating addition of illegitimate nodes in public blockchains. The proposed algorithm exploits the properties of modular inverses of integers in such a way so as to create a tradeoff between computational complexity and storage space required. The values of the parameters can be adjusted to set the difficulty level of corona graph according to application requirements ranging from highly critical to common client server applications. We provide formal proof of correctness of our approach and demonstrate analytically that the proposed technique is resilient and offers a practical solution for identifying adversaries. </p>Shalini Agarwal
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s450455An Identification and Analysis of Harmful URLs through the Application of Machine Learning Techniques
https://ijisae.org/index.php/IJISAE/article/view/4905
<p>Malicious URLs pose a significant cyber security threat, posing risks to user security and causing substantial financial losses. Traditional detection methods relying on blacklists are limited in addressing rapidly evolving threats. As a response, machine learning approaches have gained popularity for enhancing the efficiency of malicious URL detection. This paper presents a detailed analysis, offering a structured insight into various aspects and formally defining the machine learning task of identifying malicious URLs. It delves into feature representation, algorithm design. The objective of survey is to provide a detailed analysis of harmful URLS not only to researchers but to cyber security experts.</p>Swagat M. KarveShital Kakad Swapnaja Amol Ashwini B. GavaliSonali B. Gavali Shrinivas T. Shirkande
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s456468Bitcoin Price Prediction using Twitter Sentiment Analysis
https://ijisae.org/index.php/IJISAE/article/view/4906
<p>This work explores predicting Bitcoin prices using sentiment analysis on Twitter. Leveraging machine learning and rule-based methods, the study correlates social media sentiment, especially on Twitter, with dynamic Bitcoin prices. The dataset combines historical Bitcoin prices from Yahoo Finance and relevant Twitter data. Preprocessing involves cleaning tweets, calculating sentiment scores, and merging datasets. Results indicate that SGD regression and ridge regression achieve the best performance with a Validation-MAPE of 8.45%. Despite Bitcoin's volatility, the study highlights the potential of sentiment analysis in forecasting values, shedding light on the intricate relationship between social media sentiments and cryptocurrency markets.</p>Himanshu ButeAbhyuday SinghShlok NandurbarkarShivali Amit WaglePreksha Pareek
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s469477A Practical Approach to Software Cost Estimation Using Stochastic Modelling
https://ijisae.org/index.php/IJISAE/article/view/4907
<p>Software cost estimation plays a very critical role in Software Project Management. If the cost of the software has not been estimated properly, it can have a drastic impact on the project execution and delivery. Traditional models for software cost estimation fail to model correctly, the cost components associated with the project. There is enormous research literature related to software cost estimation but only a handful of them relate to cost measures that include both software development and software support. This is mainly because, in recent years, there has been a remarkable change in the way the software is now developed and supported. Needless to say, the software exists everywhere from elementary education to nuclear reactors and from civil engineering to genetic engineering. As such, one cannot bind it into the same set of measures related to development and delivery. In this research, we focus on customers like hotels, airways, banking, etc., particularly massive ERP systems, for development and customization support. Such software once purchased, requires one or more support teams to ensure its availability for the client. The infrastructure teams usually maintain server support whereas the application maintenance teams provide customization and functionality support. In this research, we have developed a model that considers the fixed cost and the recurring costs associated with the software. The fixed cost is related to the cost of development whereas the recurring cost involves the cost associated with the cost of cloud/on-premise deployment and the cost associated with support teams. The contribution of this research is twofold. We have proposed a model that considers the largest set of parameters of cost-related estimation, aligned with both development and support, which is highly mapped to ERP-like software. To the best of our knowledge and belief, no existing research considers all these parameters. To make the analysis applicable to a number of case studies, we have fuzzified the parameters to make them align with linguistic hedges. The possible deviations in the cost computation are estimated using linear regression ML models. We have considered the supports and customization part in accordance with modern bug-tracking tools like JIRA. The analysis is done for the case of an educational ERP with LMS and compared the result with those available as open-source in the UCI repository.</p>Swati SaxenaShiv Kumar SinghNargish GuptaMeena MalikAnkur Goyal
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s478488Applications and Use Cases of Millimeter Wave Communications in 5 G
https://ijisae.org/index.php/IJISAE/article/view/4908
<p>Due to the 5th Generation (5G) mobile communication system's 10 Gbps data throughput and around 1 ms latency, there has been a tremendous rise in data traffic. The actual 3 GHz radio band is become so congested as cellular data demand rises. This prompts the search for newly assigned mobile communication frequency bands that can provide a wide range of spectrum. The use of millimeter wave (mm-wave) can overcome this. In order to support future multi-gigabit-per-second mobility, image, and multimedia applications, mm-wave communications aim to make use of the vast and underutilized bandwidth. Although millimeter wave technology has been around for a while, it has mostly been used in military application<strong>. </strong>The academic community, corporate community, and standards body have significantly boosted mm-Wave technology due to the development of process technologies and low-cost integration solutions. This paper's major goal is to shed somehow mm-Wave can be used for fifth-generation communications and how next-generation customers can greatly profit by making ethical use of the bandwidth present in the mm-Wave spectrum, which ranges from 30GHz to 300GHz. Additionally, we analyze mmWave applications in this study to show how mmWave technology may be used to deliver various services.</p>Shivam AwasthiVivek Kumar SharmaJitendra KumawatShruti MathurSatyajeet SharmaGajanand Sharma
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s489496Overcoming Occlusion Challenges in Human Motion Detection through Advanced Deep Learning Techniques
https://ijisae.org/index.php/IJISAE/article/view/4909
<p>Researchers engaged in Human Motion Detection (HMD) grapple with a primary challenge related to occlusion, where individuals or their body parts become obscured within image or video frames. Occlusion manifests in two distinct forms: Self-occlusion, occurring when one part of the human body hides another, and External occlusion, arising from external objects obstructing humans. This proposed work specifically focuses on self-occlusion and partial-occlusion. To discern human motion from visual data, three fundamental methods are deployed. The initial method, motion segmentation, entails identifying the moving object in a video. The second method, Object Classification, determines whether the moving object is human. The final method, the Tracking algorithm, is employed for identifying human gestures. Occlusion persists as a central concern in HMD. In our proposed methodology, we employ a Mask Region-based Convolutional Neural Network (Mask R-CNN) for motion segmentation to address the occlusion challenge. Object classification utilizes a Recurrent Neural Network (RNN), and for tracking human motion, even during self-occlusion, Multiple Hypothesis Tracking (MHT) is applied. This study presented an innovative hybrid algorithm, the Whale Optimization Algorithm and Red Deer Algorithm (WOA-RDA), demonstrating superior convergence speed coupled with high accuracy. Our HMD approach incorporates an RNN trained with 2D representations of 3D skeletal motion. Diverse datasets, encompassing scenarios with and without occlusion, are integrated into our proposed work. The experimental findings underscore the effectiveness of our approach in accurately identifying human motion under varied conditions, including both with and without occlusion scenarios.</p>Jeba Nega ChelthaChirag SharmaPankaj DadheechDinesh Goyal
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s497513Advancing Gastrointestinal Disease Detection through Artificial Intelligence: A Comprehensive Analysis
https://ijisae.org/index.php/IJISAE/article/view/4910
<p>In the quest to enhance the precision of gastrointestinal disease detection, Artificial Intelligence (AI) emerges as a beacon of hope, offering new perspectives in a field where accuracy can mean the difference between life and death. This study delves into the transformative role of AI in diagnosing gastrointestinal ailments, a domain where traditional methods often grapple with challenges of accuracy and early detection. With gastrointestinal disorders affecting a significant portion of the global population and being a leading cause of mortality and morbidity, the urgency for more efficient diagnostic tools is paramount. Recent advancements in AI, particularly in deep learning, have shown promising results in interpreting complex medical images, a task that has historically been reliant on the subjective expertise of clinicians. Our research navigates through these advancements, critically analyzing the efficacy of AI in identifying a range of gastrointestinal diseases from various imaging modalities. We meticulously examine case studies and current applications where AI has successfully aided in disease detection, contrasting these AI-driven methods with traditional diagnostic approaches. The findings reveal a remarkable potential of AI in enhancing diagnostic accuracy, while also highlighting some of the current limitations and areas needing further exploration. This study, grounded in recent real-world applications and data, aims to shed light on the potential of AI as a tool not just for augmenting medical diagnostics but also for revolutionizing patient outcomes in gastrointestinal healthcare.</p>Rakesh SharmaC. S. Lamba
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s514518Modified Fuzzy C-Means Clustering for Anomaly Detection in Bio-medical Data.
https://ijisae.org/index.php/IJISAE/article/view/4917
<p>Bio-medical data for different diseases are always with noise and outliers. As the source and medium differs from place to place and time to time it happens to be noisy. In this work, the authors have tried to analyse biomedical data statistically using principal component analysis. Here, the fuzzy centroid is modified with opposition learning based algorithm. Due to optimal algorithms, the modified fuzzy c-means utilized for clustering that performs excellent in terms of outlier detection. The data taken from UCI machine learning repository are of classification type. It is shown in the result section that the outliers have been detected successfully.</p>Srikanta Kumar SahooPriyabrata PattanaikMihir Narayan Mohanty
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s519526BMIRTE: Design of a Bioinspired Model for Improving Readability of Translated Sentences via Ensemble Operations
https://ijisae.org/index.php/IJISAE/article/view/4918
<p>Sentence readability is determined via multiple metrics that include, Flesch Reading Ease, Fog Scale, Flesch-Kincaid Grade Levels, Smog Index, Coleman-Liau Index, Automated Readability Index, Dale-Chall Readability Score, Linear Write Formula, and their consensus. But individual use of these models results in uncertain sentence structures, which limits their usability levels. Moreover, scanning through every combination of these techniques to generate fused readability models is impractical and highly complex under real-time scenarios. To overcome these limitations, this text proposes design of a novel Bioinspired Model for Improving Readability of Translated Sentences via Ensemble Operations. The proposed model initially collects a set of translated texts, and applies stochastic ensemble readability testing via Genetic Algorithm (GA) based process. Due to use of stochastic modelling, the proposed optimizer is capable of identifying corpus specific readability evaluation techniques, that can be used to improve overall readability of multiple sentence types. To perform this task, a readability-based fitness function was evaluated, which assisted in identification of optimum ensemble operations. The model also tracks iterative performance levels of different ensemble combinations, which assists in incrementally improving real-time readability performance for different corpus types.The proposed model was evaluated on multiple translated corpuses, and it was observed that the proposed model outperformed various state-of-the-art methods in terms of readability accuracy, precision, recall, computational delay and memory requirement metrics. Due to which, it was observed to be capable of deployment for a wide variety of real-time post-processing scenarios for translated-texts.</p>Pooja P. WalkeFarha Haneef
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s527535A Pragmatic Review of Learning Models Used for Unsupervised Analysis of Existing Cyber Physical Deployments from an Empirical Perspective
https://ijisae.org/index.php/IJISAE/article/view/4919
<p>Cyber-physical deployments include game engines, multimedia systems, internet of Things (IoT) systems, etc. Each of these models has certain inputs, several processing layers, and certain outputs. Monitoring & control of such deployments can be automated via their unsupervised analysis, which requires deep learning & pattern analysis methods. A wide variety of such models are proposed by researchers and system designers, but each of them has its own nuances, advantages, limitations, & future research scopes. Moreover, these models have different performance characteristics, that vary in terms of analysis accuracy, precision, recall, fMeasure, delay of analysis, response time, computational complexity, etc. Thus, while deploying such learning models, researchers & system designers are required to perform manual analysis, validation, and testing for automation & control. Due to this cumbersome process, the cost & time to market for these unsupervised control models is very high, which limits their scalability, and deployment capabilities. To overcome this issue, a detailed characteristic discussion of these models is done in this text. Based on this discussion, researchers will be able to identify existing unsupervised & semi-supervised learning models, which closely match their deployments. These models are further analyzed in terms of their performance metrics, that includes, accuracy of analysis, response time needed for control, delay needed for analysis, precision of analysis, computational complexity, and cost of deployment. Using these metrics, researchers can evaluate best performing models for their deployments, which will assist them in reducing cost, and time needed for automating their cyber physical systems. This text also discusses certain future prospects that can be explored by researchers in order to further enhance quality of their deployments.</p>Varsha Haridas Sadrani G. V. V. Jagannadha Rao
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s536552Strengthening AI Governance through Advanced Cryptographic Techniques
https://ijisae.org/index.php/IJISAE/article/view/4920
<p>This research elucidates the pivotal role of advanced cryptographic techniques in fortifying the governance of artificial intelligence (AI) systems. Addressing the escalating challenges of accountability, transparency, and ethical AI development, the study explores the application of cryptography to enhance AI technologies' security, privacy, and accountability. The manuscript offers practical insights into cryptographic solutions, demonstrating their efficacy in mitigating risks and fostering responsible AI by combining a thorough literature review with empirical evidence. The findings contribute valuable perspectives for policymakers, practitioners, and researchers seeking to establish robust governance frameworks for the ethical deployment of AI technologies.</p>Alok KumarUtsav UpadhyayGajanand SharmaRavi Shankar SharmaNeha MishraJitendra Kumawat
Copyright (c) 2024 Alok Kumar, Utsav Upadhyay, Gajanand Sharma, Ravi Shankar Sharma, Neha Mishra, Jitendra Kumawat
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s553560An Approach for Product Recommendation using Light GBM
https://ijisae.org/index.php/IJISAE/article/view/4921
<p>Attracting clients is the main task of online e-commerce websites. Systems for providing recommendations are essential for engaging clients. Customer reviews play a crucial role in analyzing the product. Product insights can be provided by sentiment analysis of customer reviews. Websites routinely recommend products despite bad user reviews, which dissatisfy customers. Hence there is a need for a more accurate model recommending the products. In this work, a machine learning model is proposed that suggests a product with a greater user sentiment for positivity. Models are developed to analyze the sentiment of product reviews using the algorithms ADABoost, Light GBM, Gradient Boosting, Extreme Gradient Booting, and Extreme Gradient Boosting coupled with Random Forest. Based on the performance of the models, the Light GBM model is considered for building the product recommendation system. The proposed model gave better results when compared to existing models.</p>I. S. Siva RaoParasa Rajya LakshmiDasari N. V. Syma KumarAkkala Yugandhara ReddyJayavarapu KarthikBadugu Bhavana
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s561570Developing A Framework for Diseases of Banana Plant Based on the Deficiencies of Minerals in the Soil.
https://ijisae.org/index.php/IJISAE/article/view/4922
<p>Banana cultivation is of significant economic and nutritional importance worldwide. However, the growth and health of banana plants are heavily reliant on the mineral composition in which they are cultivated. This study presents a comprehensive framework for diagnosing and mitigating of banana plants diseases through an analysis of soil mineral deficiencies. The primary objective regarding to the research is to establish a framework for effective disease management in banana plants by considering the role of soil mineral deficiencies. Specifically, our goal is to: Investigate and understand the relationship between soil mineral deficiencies and the phenomenon of diseases in banana plants. Identify common banana plant diseases associated with specific mineral deficiencies. Develop predictive models and algorithms that make work of machine learning techniques to forecast disease risks based on soil mineral content. Suggest practical recommendations for mitigating disease risks through soil management and targeted fertilization strategies. The background for this proposal stems from a growing concern in the agricultural community about the devastating impact of diseases on banana plantations. Historically, disease management in banana plants has been approached mostly through pest control and environmental interventions. However, as evidence linking mineral deficiencies in the soil to disease occurrence became apparent, it highlighted the need for a more holistic and proactive approach to disease management.</p>Chukka KeerthanaPeram TejasreeMudivarthi Venkata Subba RaoR. S. Sai Pavan KumarPrasanth Yalla
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s571577Predictive Modeling of Bitcoin Prices using Machine Learning Techniques
https://ijisae.org/index.php/IJISAE/article/view/4923
<p>This research paper aims to comprehensively examine diverse algorithms employed in forecasting the price dynamics of bitcoin. The study's outcomes have undergone careful analysis, shedding light on emergent trends poised to exert influence on the cryptocurrency market in the proximate horizon. Notable among the algorithms scrutinized are the K-Nearest Neighbors (KNN), Logistic Regression, Linear Regression, and Seasonal Autoregressive Integrated Moving Average (SARIMA). A brief comparison of these algorithms has been done, with the intent of identifying the ideal machine learning-based algorithm for predicting Bitcoin's price. The preeminent criterion for model selection is predicated upon achieving optimal accuracy, culminating in the recognition of Linear Regression as the most adept algorithm for precise Bitcoin price predictions.</p>Gagandeep KaurPoorva AgrawalLatika PinjarkarRutuja Rajendra PatilRupali GangardePriya ParkhiBhagyashree Hambarde
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s578586The Role of Work Engagement in Mediating Job Stress and Religiousness on Subjective Well-Being at Secondary Education Institutions in Medan City
https://ijisae.org/index.php/IJISAE/article/view/4924
<p>This study supposed to determine the role of work engagement in mediating job stress and religiosity towards subjective well-being at senior secondary education institutions in Medan City. The research method that will be used in this research is an associative method with a quantitative approach. In this study, the population and saturated samples taken were all teachers in 6 private universities in Medan City, totaling 333 respondents. The analysis model using structural equation modeling (SEM) is a collection of statistical techniques that allow testing a series of relatively complex relationships simultaneously. The results showed that directly job stress has no effect and is not significant on the work engagement of teaching staff at secondary education institutions in Medan. Directly job stress has no effect and is not significant to the subjective well being of teaching staff secondary education institutions in Medan. Directly religiosity has a significant effect on the work engagement of teaching staff at secondary education institutions in Medan. Directly work engagement has a significant effect on the subjective well being of teaching staff at secondary education institutions in Medan. Indirectly, work engagement does not moderate job stress on the subjective well being of teaching staff at secondary education institutions in Medan. Work engagement indirectly moderates religiosity on the subjective well being of teaching staff at secondary education institutions in Medan.</p>Indra KesumaZainuddin ZainuddinSofiyan SofiyanSalman Faris
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s587597Adaptive Red Deer Optimization (ADRO) Technique for Energy Efficient VM Migration in Cloud Computing
https://ijisae.org/index.php/IJISAE/article/view/4925
<p>Several migration techniques are available for migrating Virtual machine (VM) from one host to another. But they fail to consider the migration cost while determining the energy consumption during migration. The migration cost includes the migration time and distance. Hence the objective of this work is to design an optimized VM migration technique which simultaneously reduces the energy consumption and cost while avoiding (QoS) degradation. For this, Adaptive Red Deer Optimization algorithm for energy efficient VM migration (ARDO-EEM) in cloud computing is proposed. In ARDO-EEM, the overloading probability of each host is determined based on the total resource utilization of the host. Then the overloaded hosts are categorized into heavy, medium and light depending on two threshold values. VMs to be migrated are selected from the heavy and medium overloaded hosts with energy consumption higher than the available energy. The target VMs are selected using the ARDO algorithm based on the migration energy and resource utilization. Then each VM in the migration list is relocated to the selected target VM. Experimental results show that the proposed ARDO-EEM attains increased resource utilization with lesser power consumption and response delay.</p>D. KomalavalliT. Padma
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s598608Introduction to Writing for Children with A K-Nearest Neighbor Approach
https://ijisae.org/index.php/IJISAE/article/view/4926
<p>This paper presents the design, development, and evaluation of an innovative mobile application targeted at promoting writing skills among young learners. Learning to write is a basic skill for children at an early age, but acquiring this competency is often a challenging process. The primary aim of this research was to design an application that is child-friendly, interactive, and effective in teaching children how to write. The application was built with a specific focus on aiding children in the journey of learning writing skills starting from individual characters and gradually progressing to words. So that this learning can make it easier for children to understand every letter and word that they can learn to write. The application utilizes the robustness of K-Nearest Neighbor (K-NN) algorithm for the recognition of children's handwriting. The K-NN algorithm was employed as the core engine to recognize and assess the child's handwriting and provide immediate feedback. Based on the results of preliminary testing shows promising results, with improved writing skills and high engagement levels among a group of test students.</p>Rini WongsoViolitta YesmayaSuryanto WijayaArya Dwi KurniawanMochamad Harsya Joe Andaru
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s609614An Enhanced Three Layer Cryptographic Algorithm for Cloud Information Security
https://ijisae.org/index.php/IJISAE/article/view/4927
<p>The rapid adoption of cloud computing has ushered in unparalleled opportunities for efficient data storage and processing, but it also brings forth significant challenges related to information security. This study focuses on the design and analysis of effective techniques for enhancing information security in cloud computing environments, with a particular emphasis on the hybrid encryption technique. This research delves into the theoretical underpinnings of AES+ChaCha20 and SHA-3 evaluates its suitability for cloud-based applications, considering factors such as encryption strength, computational efficiency, and resistance to cryptographic attacks. The SHA-3 algorithm is employed to generate fixed-size hash values from input data. In this study, the plain text file is encrypted using the AES+ChaCha20, the cipher text file is hashed using the SHA3 algorithm to produce a message digest. Through a comprehensive analysis, including performance benchmarks and security assessments, this study aims to provide a nuanced understanding of the effectiveness of AES+ChaCha20 in mitigating common threats in cloud computing environments. The outcomes of this research contribute valuable insights to the on-going discourse on information security in cloud computing, offering a foundation for the development and implementation of robust security measures. As organizations increasingly rely on cloud services, the findings of this study are poised to inform best practices for securing data and ensuring the confidentiality and integrity of information in cloud-based systems.</p>Anjana AnjanaAjit Singh
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s615627Millimetre-wave Massive MIMO Systems with Optimal Hybrid Precoding
https://ijisae.org/index.php/IJISAE/article/view/4928
<p>For millimetre-wave (mm-wave) massive-MIMO systems, unrestricted digital precoding is problematic due to the requirement for separate radiofrequency (RF) chains for each antenna, which results in significant costs and power consumption. Hybrid precoding, which enhances data stream adaptability and permits large MIMO transmission in mm-wave communications, is introduced as a workable approach in this study. Our study focuses on capacity analysis and hybrid precoding optimization in mm-wave massive MIMO systems, opening the path for next-generation wireless communication technologies that are more effective, affordable and have high throughput and low energy consumption.</p>Shailender ShailenderShelej KheraSajjan Singh
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s628635Machine Learning-Based Real-Time Fault Prediction: Enhancing Distribution Transformer Health Monitoring System
https://ijisae.org/index.php/IJISAE/article/view/4929
<p>Addressing the critical concern of real-time monitoring for transformers to mitigate potential operational problems due to damages, this paper highlights the substantial costs linked with maintenance and replacement, posing significant challenges. To address this, an IoT-based monitoring system is devised, ensuring continuous health assessment by tracking Voltage, Current, Temperature, and load capacity. The collected data is sent for analysis to a central server, offering insights into the broader electrical system's performance. IoT integration strengthens security, provides accurate environmental insights, and facilitates early fault detection, enabling prompt repairs and minimizing system failures. In contrast, traditional manual monitoring struggles to detect subtle changes, while IoT-driven remote monitoring requires a robust centralized data infrastructure and real-time transmission, preventing major faults and ensuring equipment protection. This approach reduces risks through centralized remote transformer data collection, complemented by machine learning techniques for proactive flaw prediction.</p>Deepak KulkarniN. Kumar Swamy
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s636643Enhancing Financial Insights: Integration of various Machine Learning Techniques
https://ijisae.org/index.php/IJISAE/article/view/4930
<p>The convergence of machine learning has catalyzed a paradigm shift in the financial realm, empowering institutions to glean deeper insights and make informed decisions. This abstract explores the multifaceted integration of these technologies, unveiling their impact on financial operations, risk management, predictive analytics, and customer-centric services. By harnessing vast datasets and leveraging sophisticated algorithms, this fusion enables proactive risk assessment, precise predictive models, and personalized financial strategies. However, while revolutionizing the sector, it poses challenges in ethical use, data privacy, and interpretability. This studydelves into the transformative potential and the accompanying considerations in the synthesis of machine learning within the financial domain.</p>S. N. GunjalS. ShiyamalaPriyanka D. HalleAnuradha S DeshpandeMinal Vilas GadeTushar Jadhav
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s644650Incorporating Machine Learning into Environmental Impact Assessments for Sustainable Development
https://ijisae.org/index.php/IJISAE/article/view/4931
<p>The growing concerns surrounding environmental degradation and the imperative for sustainable development have brought about a significant paradigm shift in the methodologies employed in Environmental Impact Assessments (EIAs). This research paper investigates the application of Machine Learning (ML) methodologies to Environmental Impact Assessments (EIAs) to improve their precision, productivity, and overall efficacy in the pursuit of sustainable development. By conducting an extensive review of pertinent scholarly works, case studies, and emergent patterns, the objective of this paper is to clarify the possible advantages and obstacles that may arise from the integration of machine learning into the EIA procedure. The subtopics that have been identified encompass the preprocessing of data predictive modeling, decision support systems, and the ethical implications that arise from the convergence of technology and environmental preservation. In conclusion, this study proposes that environmental science and state-of-the-art ML methodologies work in tandem to foster a more sustainable and resilient future through harmonious collaboration.</p>Obulesu VarikuntaA. Sarveswara ReddyK. Sathish
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s651656"Innovative Insights: Unleashing Machine Learning for Precise COVID-19 CT Scan Diagnosis"
https://ijisae.org/index.php/IJISAE/article/view/4932
<p>Effectively managing and mitigating the impact of COVID-19 requires swift and accurate identification of cases. This study explored the use of CT scan images to diagnose COVID-19 infections and evaluated the effectiveness of various approaches based on deep learning and machine learning methodologies. LSTM, ResNet50, MobileNet, KNN, SVM, decision tree, Nave Bayes, logistic regression, CNN, and LSTM were used for training and evaluation. Measures, including precision, accuracy, recall, F score, and false prediction rate, are computed using CT scan image collection and preprocessing. Our results show the remarkable performance of the CNN algorithm, which achieved 100% accuracy, 100% recall, 100% F score, 100% precision, and a 0% false prediction rate. Based on the comparison analysis, both the KNN and SVM algorithms showed promising results. These results suggest that COVID-19 patients can be reliably identified from CT images using deep learning and machine learning techniques. Subsequent investigations should explore transfer learning methodologies and ensemble models and amalgamate many modalities to enhance the algorithms' overall applicability and precision in diagnosis.</p>Kumar KeshamoniL. Koteswara RaoD. Subba Rao
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s657665AI-Based Prediction of Myocardial Infarction in Patients Using Various Algorithms
https://ijisae.org/index.php/IJISAE/article/view/4933
<p>Myocardial Infarction (MI) is among the primary causes of mortality worldwide and early detection of this condition can improve patient outcomes. Artificial Intelligence (AI) have shown promise in predicting Myocardial Infarction in patients, but the optimal algorithm for this task is not yet clear. This study assessed the efficacy of four ML algorithms - K-nearest neighbours (KNN), Logistic regression, Support Vector Machine (SVM), and random forest analysis - in predicting MI in patients. This study includes Myocardial Infarction dataset of 303 patients with details of medical history, demographic info, as well as clinical constraints. The data pre-processing was done with missing values, removing outliers and normalizing the data. In addition, feature selection approaches identify the most relevant predictors of myocardial infarction. The accuracy metrics are determined by evaluating the training performance of the four algorithms on a practice set. When the results are compared, Logistic Regression outplays the others with an overall accuracy of 81.32%. However, K-nearest neighbors, SVM, and Random Forest had accuracy rates of 65.93%, 54.95%, and 81.32%, respectively. Thus, according to our research findings, Logistic Regression is the optimal algorithm for predicting MI in patients. It is a straightforward, interpretable, and efficient technique that can be used in clinical decision-making. Our findings provide essential data about the use of machine learning algorithms to predict myocardial infarction and can help guide future studies in this area.</p>Shridevi K. JamageRamesh Y. MaliVirendra V. Shete
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s666673Security in Cloud Computing: Issues and Challenges
https://ijisae.org/index.php/IJISAE/article/view/4935
<p>Cloud computing is a popular social phenomena that most people use on a regular basis. There are some themes that are universally accepted, much like at any big social event. In the current context, cloud computing is viewed as a quickly evolving field that, with the aid of hardware and software virtualization, may instantly provide extensible services via the Internet. The ability to lease and release resources on a flexible basis in accordance with user requirements is the largest benefit of cloud computing. Additional advantages include increased efficiency, which balances operating costs. This lowers the expensive price of software and hardware. Adopting the newest cloud technology has numerous advantages, but there are privacy concerns as well because data on the cloud might move from provider to provider at any time.</p> <p>The past few decades have seen a steady increase in demand, which has resulted in a notable rise in interest in cloud computing. Organisations using cloud-based data storage solutions may benefit from a number of factors. These include the potential financial benefits from cloud computing, simpler IT infrastructure administration, and remote information access from any location in the globe with a stable Internet connection. Further research is needed to fully understand the security and privacy issues related to cloud computing. In earlier investigations, researchers from standards organisations, academia, and industry have suggested possible remedies for these problems.</p> <p>Most issues stem from the fact that the consumer no longer has control over their data because it is stored on a single machine that is owned by the cloud provider. Because their interests may differ (the user may wish his information to be kept private, but the owner may desire to utilise this for his own company), this happens when the owner of the remote server is a different person or organisation than the user. Concerns regarding automated management, guaranteed IoS provisioning, and uncertainty about future cloud system upgrades further impede the adoption of cloud technology. This study covered the fundamentals of cloud computing as well as security flaws, vulnerabilities, and fixes. The article also discusses a number of important cloud-related subjects, including cloud technologies, cloud architectural frameworks, and cloud security ideas, risks, and threats.</p>Mohammad Ahmar KhanPratibha GuptaAbdulsatar Abduljabbar SultanPreeti SinghShivam ShivamMelanie Lourens
Copyright (c) 2024 Mohammad Ahmar Khan, Pratibha Gupta, Abdulsatar Abduljabbar Sultan, Preeti Singh, Shivam, Melanie Lourens
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s674681Performance of Flood Early Detection System (FEDS) and Artificial Neural Network on Predicting Flood in the City
https://ijisae.org/index.php/IJISAE/article/view/4936
<p>An integrated flood detection system that is easily accessible to the public is one of the efforts to reduce the impact of flooding in flood-prone locations. The performance of flood detectors integrated with the internet helps make it easy Public for access information about possibility happening flood. Information about bulk rain, high water level, the water discharge will Becomes indication possibility happening downstream flooding. Tool prototype detector flood this be equipped with temperature and humidity sensor as addition information for society. Research results show that performance tool detector flood already good because capable give information related to data that can be made indication possibility happening flood. Sensors used have score small mistake and after calibrated got score constant for standardize results testing. Bulk sensor test results rain produces an average error of 3% and after calibration obtained constant of 1.03. High sensor test results water level has an average error of 1.07% and after calibrated obtained score constant of 0.98. Humidity sensor testing have the average error value is 3% and after calibration obtained score constant 1.03.</p>Agung Wahyudi BiantoroS. I. Wahyudi Moh. Faiqun Ni'amSubekti Subekti
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s682688Enhancing Oral Squamous Cell Carcinoma Detection: A Transfer Learning Perspective on Histopathological Analysis Using ResNet-18, AlexNet, DenseNet-169, and DenseNet-201 with Cyclic Learning Rate
https://ijisae.org/index.php/IJISAE/article/view/4937
<p>: In this study, an innovative method is introduced for the early identification of Oral Squamous Cell Carcinoma (OSCC) by employing deep learning techniques to analyze histopathological samples. Four prominent neural network architectures, ResNet-18, AlexNet, DenseNet-169, and DenseNet-201, are utilized to scrutinize biopsy specimens for cancerous anomalies. The approach incorporates Cyclic Learning Rate (CLR) for dynamic adaptation of learning rates during the model's training. ResNet-18 benefits from skip connections to enhance gradient flow, while AlexNet and DenseNet architectures significantly contribute to precise image categorization. DenseNet's distinctive feature reuse mechanism effectively mitigates the vanishing gradient issue. The research underscores the potential of deep learning in enhancing early OSCC detection, offering a promising avenue for more efficient cancer screening and treatment.</p>Aarti YadavSurendra Yadav
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s689699SFT for Improved Text-to-SQL Translation
https://ijisae.org/index.php/IJISAE/article/view/4938
<p>Large Language Models (LLMs) have proved significant proficiency when comes to code generation especially in Structured Query Language (SQL) for databases and recent successful Text-to-SQL method involves fine-tuning pre-trained LLMs for SQL generation tasks. Transforming natural language text into SQL queries, has been attempted to solve with various learning techniques including Few-shot learning[1], fine tuning. In this paper we propose Supervised fine-tuning (SFT) as a better alternative for learning technique for text-to-SQL generation task using Code-Llama that pushes state of art accuracy on spider test suite to 89.6% on dev set which represent first instance of surpassing the earlier best-in-class with 5.5% higher score and 86.8% of exact match accuracy on dev set. Furthermore, we demonstrate that properly prompted LLM along with SFT provides far fewer hallucinations and much more robust LLM that can be used as a general tool for any text-to-SQL generation use case.</p>Puneet Kumar OjhaAbhishek GautamAnkit AgrahariParikshit Singh
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s700705Applying Multi-YOLO for Enhanced Product and Fire Detection in Image Analysis
https://ijisae.org/index.php/IJISAE/article/view/4939
<p>With its wide range of uses and intense research interest, computer vision presents a challenging problem when it comes to product and fire detection in images. In addition to providing useful applications including improving consumer product information, enabling image-based rapid payments, automating product availability management, and building early fire warning systems, this task entails identifying goods and fire in photographs with diverse backdrops. However, there is a problem with the widely held belief in product detection research, which holds that training data should reflect actual situations. The effectiveness of product detection systems is impacted by the fact that testing data obtained in a variety of contexts does not match training data, which is frequently gathered under perfect conditions. This work presents a deep learning method for image-based product detection in response to these difficulties. To identify products in photos, the suggested model, known as Multi-YOLO, makes use of several YOLO models. Every element operates as a separate YOLO model, and Fusion rules combine their outputs to create a single output. The experimental results show how well the suggested model works, especially when applied to our collection of product photos, and emphasize its potential for reliable product detection in practical settings. Furthermore, the study's integration of the Multi-YOLO model within a comprehensive early fire alert system paves the way for enhanced fire prevention strategies and improved public safety outcomes.</p>Hai Tran Son
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s706714EEG Signal Analysis for Epilepsy Patterns Using EDFBrowser
https://ijisae.org/index.php/IJISAE/article/view/4940
<p><strong>Background- </strong>In the healthcare system, biomedical signals are significant. It aids in diagnosing medical conditions and offers valuable information about a patient's health. Neurologists can find meaningful information and patterns by carefully examining electroencephalogram (EEG) data, which enables sufferers to receive proper care at the right time.</p> <p><strong>Objective- </strong>Biomedical EEG signals can be used to study the neurological condition known as epilepsy. Signals from an EEG can be used to study brain activity and identify neurological disorders. Recurrent seizures are a sign of epilepsy that affects the brain.</p> <p><strong>Methods- </strong>A medical history, neurological examination, and diagnostic EEG tests are frequently used to identify seizures. EEG bio-signals can be viewed and analyzed using an open-source software program called an EDFBrowser.</p> <p><strong>Results- </strong>This research defines the various ways, including visual inspection, frequency analysis, time-frequency analysis, and spike detection, to read EDF (European Data Format) for epileptic patterns. Also, to define the amplitude-frequency relationship, analyze EEG signals, and examine brain activity in different frequency bands, like Delta, theta, alpha, beta, and gamma, employing Fast Fourier Transform (FFT).</p> <p><strong>Conclusion-</strong> In neuroscience, it must correctly interpret the EEG signal patterns to analyze the condition of the brain.</p>Ashish SharmaVinai K. Singh
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s715723No Reference Quality Assessment Metric for Multi-spectral and Multi-Modal Image Fusion using Sparse Approximate Variational Autoencoder
https://ijisae.org/index.php/IJISAE/article/view/4941
<p>Unlike natural image quality assessment approaches, satellite stereo images have various quality criteria in different application contexts, making it difficult to develop an appropriate objective evaluation model. The area of perceptual quality evaluation has evolved significantly and continues to expand. In the low-level computer vision field, no reference image quality assessment (NRIQA) is critical. Deep neural networks are gaining popularity for NRIQA applications. Existing deep learning-based systems are generally supervised and depend on an unrealistically huge number of labelled training data. Model-based techniques are unsupervised and flexible, but they depend on handmade priors. The majority of extant No reference image quality assessment (NR-IQA) models were designed for synthetically distorted images, however they perform badly on in-the-wild images, which are frequently used in a variety of practical applications. Blind Image Quality Evaluation Metric for Multi-spectral and Multi-modal Image Fusion Techniques is developed in this research. This No reference quality measure is examined and compared to numerous well-known cutting-edge methods and mean opinion score. The proposed quality evaluation regression models successfully predict quality score. When compared to the MOS score, archives score with 96% similarity. The suggested approach has a Pearson correlation value of 0.96 and a Spearman's rank correlation coefficient of 0.83.To use the abundant self-supervisory information and decrease the model's uncertainty, we impose self-consistency between the outputs of our quality assessment model for each image and its sparse code book. Our results demonstrate that our suggested technique outperforms other methods on Fused image datasets with distorted images.</p>Milind S. PatilPradip B. Mane
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s724732A Novel Cluster Based Video Object Segmentation for Key Frame Extraction
https://ijisae.org/index.php/IJISAE/article/view/5017
<p>Video key frames are the abstraction of content rich frames of a shot or a video that best reflects the nature of the whole video without redundancy. Object based key frame extraction techniques are capable of extracting key frames that are semantic. These techniques needs to extract the required objects or region through video object segmentation. The segmentation of objects is achieved by Fuzzy C-Means clustering as it distinguishes well across object boundaries. In this paper, Oppositional based Border Collie Optimization algorithm is proposed along with Gaussian Kernel FCM to optimize the centroids of clusters. The accuracy of the segmented objects are evaluated in terms of SSIM, BDE and VoI with the SBM-RGBD dataset. The resultant frames with segmented objects are compared with consecutive frames for change of pose of objects using key points features. When there is a considerable variations between two frames, one of the frames is selected as a key frame. The experimental results showed that the proposed BCOKFE technique improves the accuracy of the extracted key frames to 92% for the WEB data set.</p>Jeyapandi MarimuthuVanniappan Balamurugan
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s733743Breast Cancer Identification using Hybrid Algorithm
https://ijisae.org/index.php/IJISAE/article/view/5018
<p>The Soft Computing has the prospective to predict diseases based on features buried in data. Incidence and humanity rates from breast cancer have risen steadily during the previous three eras. In 2023 it is estimated 3,00,590 people were diagnosed with breast cancer. Around 2,97,790 new cases are diagnosed in women at every month. By 2030, experts predict that the annual number of new cases analysed will have reached 2.7 million with 0.87 million deaths. This breast cancer caused by many factors like various Clinical, Social, Lifestyle and Economic. So key challenge of predicting the breast cancer is the construction of prototype for addressing all notorious risks factors. The feature extraction will improve the predictive performance of a model with Convolutional Neural Network (CNN). This will retain a new recognition task based on existing network with trained weights. In addition, this model will improve the quality of extraction so it makes the best choice for analysis. In this article, hybrid method Convolutional Neural Network (CNN) with Deep feature extraction method i.e., Soft Convolutional Grad-CAM (SCGC) method is proposed to identify the breast cancer tumor along with to know whether cancer is in nodes of lymph or spread to other parts of the body.</p>S. Vani KumariK. Usha Rani
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s744753A Novel Stochastic Gradient Descent Based Logistic Regression (SGD-LR) Framework for Customer Churn Prediction
https://ijisae.org/index.php/IJISAE/article/view/5019
<p>Customer churn is a crucial issue in any company or organization, and it describes the loss of customers as a result of them switching to competitors. When there is an opportunity to discover customer churn sooner, the organization may take steps such as providing important knowledge for keeping and boosting the client count. Deep Learning (DL) models have recently gained popularity because to their remarkable performance boost in a variety of fields. In this paper, a DL-based Customer Churn Prediction (CCP) is introduced using Stochastic Gradient Descent Based Logistic Regression (SGD-LR) with an LR classifier model. Effective categorization may be achieved by combining SGD with LR. The provided SGD-LR model is evaluated against a benchmark dataset, with the outcomes examined over a range of epochs. Furthermore, a comparison study with the outcomes of existing approaches is conducted. The implementation results demonstrated that the provided SGD-LR model outperformed the current CCP models on the same dataset.</p>Suja SundramD. PoornimaPraveenkumar G. D.C. BalakumarD. SasikalaSardor Omonov
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s754765ODBFT: An Optimal Derivative based Byzantine Fault Tolerance of Blockchain Consensus Algorithm with Smart Digital Contract for Health Monitoring System
https://ijisae.org/index.php/IJISAE/article/view/5020
<p>Blockchain technology starts with crypto currency called Bitcoin. Followed by Smart contract is called digital contract, which is defined as pieces of decentralized code. It performs self-sufficient operation are executed automatically to meet certain conditions. Compared with Bitcoin application, Blockchain technology is more powerful. Overall opportunity of Blockchain technology is increasing and applicable in industry, financial transaction and healthcare system. Consensus is a digital agreement or procedure to make a common decision or agreement in a decentralized network. Different methods of consensus are used in Blockchain environment and Bitcoin network. In decentralized environment, multiple nodes can take own decision whereas some nodes act as a malicious node or faulty node. Blockchain and the Internet of Things (IoT) are fast-growing technologies this can be easily integrated and applied in various services, especially for Health Monitoring System (HMS) applications. In smart HMS, IoT devices have the functionality to store, process, and analyze sensed data collected from end user data. Storage of data is also challenging because it must consider legitimate elements, a single point of failure, data manipulation, tampering, and security. To mitigate such problems, integrate Blockchain technology and store of patient sensed data for decentralized computation. In this research, concentrate on consensus mechanism and cognitive smart digital contract in Blockchain network. Propose a decentralized Cognitive Blockchain-based HMS (CBC-HMS). Cognitive blockchain framework is combination of Cognitive Consensus Algorithm with Design a Optimal Derivative based Byzantine Fault Tolerance (ODBFT) consensus technique for a blockchain with IoT technology.Through this research work, two Byzantine Fault Tolerance (BFT) consensus algorithms are proposed for improving the consensus process, reduces fault and improve the lifetime of the network with energy efficiency. Detailed review of PBFT, Paxos, RAFT, PoA, PoAh consensus algorithms is discussed. Also to improve the decision making skill for blockchain introduced cognitive smart digital contract which creates high potential action aggrement. Proposed ODBFT algorithm compared the faulty rate, security, scalability and throughput of consensus mechanism with existing models. Finally, the advantages and disadvantages of the consensus algorithms are compared.The results show that proposed ODBFT solves the problem of Byzantine faults and guarantees stable performance.</p>V. Sarala DeviS. Radha RammohanSheela K.V. VaidehiN. Jayashri
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s766780Optimization Strategies for Performance Enhancement of Packrat Parsers
https://ijisae.org/index.php/IJISAE/article/view/5031
<p>Memoization in computing refers to storing intermediate results and referring them when same inputs appear again instead of calculating them again. Packrat parsing is a comparatively new parsing technique based on top down approach with backtracking for parsing input which uses memoization and ensures linear time parsing. Introduced in 2002 by Bryan Ford, Packrat parsing was developed with focus on computer oriented languages. The ensured linear time parsing by packrat parsers comes at a cost of huge primary memory consumption for memoization making it impractical to implement. Here we have proposed a three way approach to optimize the use of memory required for memoization. Our implementation allocates memory for memoization dynamically based on resources available. Non linear data structure eliminates the requirement of continues blocks of free memory. Using linear time searching technique it is ensured that latency is constant even in case where higher number of intermediate results are stored. The proposed implementation is a promising approach for exploiting benefits of memoization to ensure linear time parsing while avoiding burdening the system in case where primary memory is a constraint..</p>Nikhil MangrulkarKavita SinghSagar Badhiye
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s781785Advancements in Cyclostationary Technique for Cognitive Radio: An In-Depth Examination and Evaluation
https://ijisae.org/index.php/IJISAE/article/view/5032
<p>In contemporary society, information communication has reached unprecedented levels, connecting individuals at all strata, from grassroots to the highest echelons. The integration of information technology and communication has become an intrinsic part of everyday life, catering to knowledge-based, entertainment-based, economical-based, and social-based data. As the volume of information data continues to grow, Information Communication Technology (ICT) plays a pivotal role in the lives of common individuals. However, the surge in users is met with the challenge of limited resources, specifically channel bandwidth for communication. This paper addresses the issue of communication channel utilization in the face of a rising number of users. To optimize and accommodate an increasing user base within the constraints of available channels, the proposed system introduces the Enhanced Cyclostationary technique for Cognitive Radio. This innovative approach aims to maximize the efficiency of communication channels, ensuring reliable connectivity for the expanding user population.</p>Sarita CharkhaMayur JakhetePrachi WaghmareDipamala ChaudhariKetan JataleTejaswini ZopeBhushan RathiVidur Sharma
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s786793Maternal Health Transformation: Harnessing IoMT for Advanced Risk Assessment and Monitoring
https://ijisae.org/index.php/IJISAE/article/view/5033
<p>This research introduces a groundbreaking approach to enhance maternal health in remote developing regions through IoT-enabled wearable sensing technology. Despite persistent challenges in reducing maternal and fetal mortality rates, the integration of intelligent machine learning algorithms into healthcare systems presents significant promise. The proposed system not only monitors maternal well-being but also delivers real-time health updates to both mothers and their families.</p> <p>Emphasizing the critical aspect of accuracy in technological solutions, particularly in remote developing regions, this research highlights the potential transformative impact of IoT-driven wearables on maternal and infant healthcare. The study aims to contribute valuable insights to the ongoing global efforts focused on reducing maternal and fetal mortality rates through innovative and effective technological interventions.</p>Dipali PanchalKrunal Vaghela
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s794804Revolutionizing Social Media Security: Integrating Graphic Passwords and Blockchain for Enhanced User Authentication
https://ijisae.org/index.php/IJISAE/article/view/5034
<p>The Office of Communications and Marketing oversees various social media platforms, including "Facebook, Twitter, Instagram, LinkedIn, and YouTube." Safeguarding online presence is paramount for any business, emphasizing the crucial role of social media security. Protection measures encompass thwarting targeted phishing attacks, fortifying corporate accounts against hacking, and preventing social engineering scams. Social media platforms like Facebook and Twitter offer opportunities for fraudsters to exploit, employing legitimate but fake accounts to gain trust and extract sensitive information. Scammers adeptly craft fraudulent accounts, mimicking individuals or companies, spreading malware, online attacks, and misinformation to deceive followers and company employees into revealing private data or corporate secrets .To enhance security, an envisioned end-to-end app proposes registering mobile numbers with a password consisting of zigzag images or by sliding computer-generated images to create a unique pattern. This complements the existing module facilitating easier SIM card identification. Decentralized blockchain formed by users in communication, as opposed to a private blockchain with a central authority, is proposed. Our social media platform prioritizes two-factor authentication, requiring users to log in with their email and password, and then verifying the login request with a unique one-time password sent to their phone number during the initiation of the login request.</p>Geerija LavaniaGajanand Sharma
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s805812Identifying the Dominant Features in Indonesia Smart Home Dataset by Interpreting Electrical Energy Consumption Prediction Results
https://ijisae.org/index.php/IJISAE/article/view/5086
<p>Smart Home needs convergence between Machine Learning (ML) and IoT to make predictions, which means ML becomes the optimal prediction model for prediction and Interpretation. Electrical Energy Consumption is a critical problem that needs to be predicted and interpreted. The proposed study aims to find the dominant feature for the Indonesia Smart Home Dataset and prediction using K-Nearest Neighbors (KNN) with Hyperparameters (k and Distance Algorithm). The dominant feature is interpreted using SHapley Additive exPlanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME). The experiment’s optimal prediction model is validated using error evaluation parameters such as RMSE, MSE, and MAE. The model results (k = 2 and Manhattan Distance) were obtained with RMSE = 0.158, MAE = 0.115, MSE = 0.025, and Manhattan Distance. Although LIME cannot interpret the feature as global, the dominant feature can be displayed globally using SHAP. The global interpretation SHAP result is that the "AC", "Washing Machine", "Lamp", "Water Pump", and “RiceCooker” must be reduced to reduce energy consumption. The KNN learning algorithm can build the model with (k=2 and Manhattan Distance) and SHAP model interpretation. Further research is needed to search for other hyperparameters based on search algorithms to maximize KNN performance.</p>Mochammad Haldi Widianto Alexander Agung Santoso Gunawan Yaya Heryadi Widodo Budiharto
Copyright (c) 2024
http://creativecommons.org/licenses/by-sa/4.0
2024-02-232024-02-231217s813821