A Dynamic Cluster Construction Method Based on Characteristics of Query Generation in Peer-to-Peer Networks

Due to the recent development of computer and network technologies, there has been increasing interest in peer-to-peer (P2P) networks. To reduce query messages for information retrieval in P2P networks, many conventional studies proposed methods that categorize similar data items and cluster peers that have data items in the same category. Some of them aim to construct clusters dynamically when the number of queries for similar data items exceeds the predetermined threshold. However, these methods have a critical problem that the threshold is fixed for any queries. The fixed threshold cannot perform cluster construction at appropriate timings since they depend on characteristics of query generation. In this paper, we propose a dynamic cluster construction method based on characteristics of query generation. Our method can reduce query messages by constructing dynamic clusters at more appropriate timings. The results of simulation experiments show that our method can reduce query messages compared with an existing method.