Comparative analysis based on clustering algorithms

Abstract This article summarizes and evaluates the clustering effects of commonly used algorithms on data sets with different density distributions. In this paper, circled datasets, sized Gaussian mixture datasets were designed as typical datasets. Then, K-means, clustering, DBSCAN, Agglomerative developed to evaluate performance these The results show that DBSCAN is more stable when distributions are not clear. Besides, calculates shortest distance can determine type set. Moreover, it appropriate use only a single algorithm analyze dataset. It recommended multiple clusters process dataset after preprocessing.