TY - JOUR AU - Unnam, Aswani Kumar AU - Rao, Bandla Srinivasa PY - 2023/01/16 Y2 - 2024/03/29 TI - An Extended Clusters Assessment Method with the Multi-Viewpoints for Effective Visualization of Data Partitions JF - International Journal of Intelligent Systems and Applications in Engineering JA - Int J Intell Syst Appl Eng VL - 11 IS - 1s SE - Research Article DO - UR - https://ijisae.org/index.php/IJISAE/article/view/2476 SP - 51 - 56 AB - <p>Cluster analysis is the most important for the data partitions of unlabelled data in various big data applications. It analyses the data based on similarity features of data objects. Two significant steps of the cluster analysis are as follows: assess the initial cluster tendency, and explore the data partitions. Top big data clustering techniques, such as k-means ++, single pass k-means (spkm),&nbsp; mini-batch-k-means (mbkm), and spherical k-means, effectively generate the big data clusters. However, they cannot get the initial knowledge about the clustering tendency. Estimation of the knowledge about the number of clusters is known as the clustering tendency. Various estimation methods of cluster tendency are surveyed and finally investigated that visual assessment of cluster tendency (VAT) accurately assesses the clustering tendency. Finding the accurate similarity features plays a vital role in accurately assessing clusters in the VAT algorithm. This paper proposes a novel computational similarity measure for the best assessment of big data clusters. The experiments are conducted on big synthetic and big real datasets to illustrate the proposed technique's efficiency.</p> ER -