K-AP Clustering Algorithm for Large Scale Dataset

Affinity propagation clustering algorithm is with a broad value in science and engineering because of it no need to input the number of clusters in advances, robustness and good generalization. But the algorithm needs the initial similarity (the distance between any two points) as a parameter, a lot of time and storage space is required for the calculation of similarity. It's limited to apply to cluster of the large amounts of data. To solve problem, this paper brings forward K-AP cluster algorithm which integrate k-means algorithm to AP algorithm to decrease time-consuming and space superiority. The results show the K-AP algorithm is faster than the original algorithm processing in speed, and it can cluster large amounts of data, and achieve better results.

I. Introduction

Cluster analysis is an important research topic in the field of data minin g, It not only can be used as a separate tool to find data in the database which are some differences in the distribution of information but also can be used as a preprocessing step of the algorithm of other data mining. Clustering as an important method in data mining is a discovery process that groups a set of data such that the intra-cluster similarity is maximized and the inter-cluster similarity is minimized.