A new distributed clustering algorithm based on K-means algorithm

With the ever increasing trend towards data volume growth, hardware speedup and the storage capacity of the computers, the data-mining field researches have been more than before, tempted to capture the underlying rule, knowledge, and relations hidden in the data. There exist a set of different techniques concerning the Data Mining, the most paramount of which is Data Clustering. In this technique a set of homogenous data is categorized into a set of distinct categories (clusters) based on the similarities in a group of parameters. While the trend is towards distributing the data over a set of far apart clusters, this property from one side and the high computational cost of clustering algorithms from the other side impose a necessity to use parallel and distributed algorithms in this area in order to exploit their better performance. In this paper we envision a distributed clustering algorithm which is scalable and provides cooperation while preserving a high degree of independency for each site.