Performance Analysis of Hierarchical and Non Hierarchical Clustering Techniques
Md. Siraj Ud. Doulah, Md. Azizul Hakim, Md. Abdul Hamid · 2020
Clustering is a procedure to organizing the objects in to groups or clustered together, based on the principle of minimizing the intra-class similarity. The various clustering algorithms are analyzed and compare the performance of clustering algorithms on aspect for validity indices to build the model. The aim is to judge the efficiency of different clustering techniques on wine dataset and determine the optimum algorithm. The results show that complete linkage and Average linkage based on different proximity measures were found to be the algorithm with most clear pictures for hierarchical clustering techniques. PAM and Robust k-means algorithms were considered an accurate clustering techniques. Model Based and Fuzzy c-means clustering were found to be next accurate algorithms after PAM and Robust k-means accordingly.