A New Approach to Distributed Clustering

Clustering is a process of dividing a set of data into several disjoining groups. One of the main problems in this domain is how to control and manage the data, which are naturally collected at different sites. In this paper we propose a technique to model the local clusters and send the clustering results to a central system to create a global cluster without moving the actual data. by using this approach we can achieve efficient accuracy without paying any local data transition extra cost.

A New Approach to Distributed Clustering Keywords:

data mining , distributed clustering , distributed knowledge acquisition.