A novel resource description bisect approach for clustering peers

Abstract

Clustering similar peers could benefit document retrieval in P2P systems. In this paper, we suggested a resource description based approach to cluster peers according to their topics. By combining the topic model (an extension of language model) technique and fuzzy set theory, we solve the key problems about resource generation and peer similarity calculation. Experiments performed on the standard data sets prove that our approach is more effective than the traditional method. © Springer-Verlag Berlin Heidelberg 2005.