군집 분석 방법에 대한 비교 연구

Abstract

Clustering analysis is a widely used in data mining to classify data into categories on the basis of their similarity. Its applications broadly range from pattern recognition to microarray, multimedia, bibliometrics, bioinfomatics, and astronomy. Through the decades, many clustering techniques, such as hierarchical and non-hierarchical algorithm have been developed. Recently, fast search by density peaks of clustering algorithm was presented in the science journal. In this thesis, we perform a comparative study of the performance of the fast search and the existing methods on the benchmark data sets in the literature. From computational experiments, we notice that the accuracy of the fast search is more or less sensitive to the value of parameters for the cluster centers.