Abstract: Distributed inverted index technology is used in many peer-to-peer (P2P) systems to help find rapidly document in which a given word appears. Distributed inverted index by keywords may incur significant bandwidth for executing more complicated search queries such as multiple-attribute queries. In order to reduce query overhead, KSS (keyword-set search) by Gnawali partitions the index by a set of keywords. However, a KSS index is considerably larger than a standard inverted index, since there are more word sets than there are individual words. And the insert overhead and storage overhead are obviously unacceptable for full-text search on a collection of documents even if KSS uses the distance window technology. In this paper, we extract the relationship information between query keywords from websites’ queries logs to improve performance of KSS system. Experiments results clearly demonstrated that the improved keyword-set search system based on keywords relationship (KRBKSS) is more efficient than KSS index in insert overhead and storage overhead, and a standard inverted index in terms of communication costs for query.

Key words: Peer-to-peer (P2P), Keyword-set search (KSS), Keyword relationship

Please provide your name, email address and a comment

Your Name: Affili: E-Mail Address:

Comment (Please comment in English):


Verification Code(click to change a pic):