Enhance Peer-to-Peer Search with Usage Pattern Clustering

• Main Authors: Ching-Hua Yu, 尤清華
• Other Authors: John Kar-kin Zao
• Format: Others
• Language: en_US
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/77588762569427200737

碩士 === 國立交通大學 === 資訊科學與工程研究所 === 94 === Clustering strategy is useful to improve blind search problem in peer-to-peer network. However most clustering in peer-to-peer network focus on comparing semantic content, but analyzing transaction behavior – download and upload in file sharing system – called usage pattern. Usage Pattern shows grouping behavior ascribed to users’ interest. Our work is to design a resource discovery flow with clusters constructed by usage pattern in the p2p environment, including usage frequency model (data transform work), clustering scheme, and search model. A connectivity-based distributed clustering algorithm called Asynchronous Spontaneous Clustering Scheme (ASC) is proposed to cope with the problem.