关键词: 仿射传播聚类, 主元分析, PCA-AP算法, 降维

Abstract: Affinity propagation clustering is a fast and efficient clustering algorithm.However，because of the overlap of the data information，error of clustering is biggish when it is applied to high-dimensional data.Concerning this problem，a new method combining Principal Components Analysis（PCA） and Affinity Propagation（AP） clustering is proposed.In this method，dimensionality of the original data is reduced on the premise of reserving most information of the variables.Then，AP clustering is implemented in the low-dimensional space.Thus，because the redundant information is deleted，the classification is accurate.The experimental results of the experiment explain that this method is effective.

Key words: Affinity Propagation（AP） clustering, Principal Components Analysis（PCA）, PCA-AP, dimensionality reduction