Customer Churn Behaviour Analysis Using Optimized XG-Boost Algorithm with Novel Hyperband Algorithm
Keywords:
Customer behaviour, Logistic Regression and XG-Boost models, Novel hyperband algorithmAbstract
The quick development of technology infrastructure has changed business procedures. Customer churning has become a major concern and threat to all companies as consumers have more options for goods and services. Primary contribution is the creation of a framework for predicting customer churn that helps businesses recognize those customers who are likely to experience turnover. A key component of customer retention tactics is customer churn forecasting, which enables companies to spot at-risk clients and take preventative action to lower churn rates. The suggested model offers a novel approach to customer forecasting churn by utilizing Hyperband cross-validation and the well-known gradient boosting technique XG-Boost to achieve effective hyperparameter tuning. The proposed model utilizes historical customer data, including demographic information, transactional records, and customer interactions, as input features. Because XG-Boost can handle intricate interactions and identify non-linear patterns in the data, it is used as the base classifier. By utilizing feature importance scores from XG-Boost, the model gains valuable insights into the significance of different features in predicting customer Churn. To enhance the performance of the model, Hyperband cross-validation is adopted for hyperparameter optimization. Hyperband adaptively allocates computational resources to different hyperparameter configurations, efficiently balancing exploration and exploitation. This ensures that the model converges to optimal hyperparameter settings within a limited computational Budget. The proposed model is evaluated on a real-world dataset from a telecommunications company, containing a diverse set of customers and labelled churn outcomes. Performance metrics, including accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC-ROC), are used to assess the model's effectiveness in predicting customer churn. The results demonstrate that the proposed XG-Boost model with Hyperband cross-validation outperforms traditional hyperparameter optimization methods, achieving higher accuracy and better generalization. The model's interpretability is enhanced through feature importance analysis, allowing businesses to understand the key drivers of churn and tailor their retention strategies accordingly.
Downloads
References
H. Zhang, D. Qiu, R. Wu, Y. Deng, D. Ji, and T. Li, “Novel framework for image attribute annotation with gene selection XGBoost algorithm and relative attribute model,” Appl. Soft Comput. J., vol. 80, pp. 57–79, 2019, doi: 10.1016/j.asoc.2019.03.017.
P. Rao and J. Manikandan, “Design and evaluation of logistic regression model for pattern recognition systems,” 2016 IEEE Annu. India Conf. INDICON 2016, no. 6, 2017, doi: 10.1109/INDICON.2016.7839010.
A. Ullah, M. I. Mohmand, H. Hussain, S. Johar, and I. Khan, “Customer Analysis Using Machine Learning-Based,” 2023.
W. Zhang and M. Wang, “An improved deep forest model for prediction of e-commerce consumers’ repurchase behavior,” PLoS One, vol. 16, no. 9 September, pp. 1–16, 2021, doi: 10.1371/journal.pone.0255906.
S. Putatunda and K. Rama, “A Modified Bayesian Optimization based Hyper-Parameter Tuning Approach for Extreme Gradient Boosting,” 2019 15th Int. Conf. Inf. Process. Internet Things, ICINPRO 2019 - Proc., pp. 22–27, 2019, doi: 10.1109/ICInPro47689.2019.9092025.
Y. Suh, Machine learning based customer churn prediction in home appliance rental business. Springer International Publishing, 2023.
V. Shrirame, J. Sabade, H. Soneta, and M. Vijayalakshmi, “Consumer Behavior Analytics using Machine Learning Algorithms,” Proc. CONECCT 2020 - 6th IEEE Int. Conf. Electron. Comput. Commun. Technol., no. Cd, pp. 5–10, 2020, doi: 10.1109/CONECCT50063.2020.9198562.
Z. Hu, Consumer Behavior Prediction Based on Machine Learning Scenarios. Atlantis Press International BV, 2023.
A. Nikitenko, I. Andersone, A. Zujevs, E. Gaile-Sarkane, and V. Bergs, “Customer behaviour forecast using deep learning networks,” ACM Int. Conf. Proceeding Ser., pp. 199–208, 2022, doi: 10.1145/3529399.3529431.
F. Safara, “A Computational Model to Predict Consumer Behaviour During COVID-19 Pandemic,” Comput. Econ., vol. 59, no. 4, pp. 1525–1538, 2022, doi: 10.1007/s10614-020-10069-3.
R. Atre and N. Tapaswi, “A Prediction of Customer Behavior using Logistic Regression, Naivesbayes Algorithm,” Int. J. Comput. Appl., vol. 183, no. 50, pp. 31–35, 2022, doi: 10.5120/ijca2022921904.
S. Sharma and A. Waoo, “Customer Behavior Analysis in E-Commerce using Machine Learning Approach : A Survey,” Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol., vol. 9, no. 1312, pp. 163–170, 2023.
A. Apichottanakul, M. Goto, K. Piewthongngam, and S. Pathumnakul, “Customer behaviour analysis based on buying-data sparsity for multi-category products in pork industry: A hybrid approach,” Cogent Eng., vol. 8, no. 1, 2021, doi: 10.1080/23311916.2020.1865598.
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
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.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. 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.