Enhanced distributed document clustering algorithm using different similarity measures

Many of the distributed environments like internets, intranets, local area networks and wireless networks have different distributed data sources. Inorder to analyze and monitor these distributed data sources specialized data mining technologies for distributed applications are required. A variety of distributed document clustering algorithms exists for this purpose. This paper presents an Enhanced Distributed Algorithm (EDA) for document clustering. This paper presents the performance analysis of the algorithm using different similarity measures like cosine similarity, Jaccard and Pearson coefficient. The test was performed on standard document corpora like 20NG (News Group), Reuters, WebKB. The performance of this proposed EDA algorithm is also evaluated using different performance factors in order to determine its accuracy and clustering quality.