Zhou Bing,Shen Junyi,Peng Qinke
(School of Electronics and Information Engineering,Xi'an Jiaotong University,Xi'an ¡¡710049,China)
Abstract:Based on the analysis of the deficiency of BIRCH algorithm, a new clustering algorithm named CLAP(Clustering algorithm based on rAndomª²samPling and cluster-feature) is proposed. CLAP preprocesses some of the data extracted from the database by random sampling technique, which decreases the running time greatly. CLAP improves the precision of clustering by setting up the diameter of the leaf nodes in the index tree and the diameter of the clusters. CLAP combines the global searching and local searching such that eliminates the influence of input order on the clustering quality. Experiment results show that the CLAP algorithm not only increases the clustering speed, but also improves the clustering quality.
Keywords:clustering£»BIRCH algorithm£»random sampling