SMBSRP: a search mechanism based on interest similarity, query relevance and distance prediction

  • Authors:
  • Fen Wang;Changsheng Xie;Hong Liang;Xiaotao Huang

  • Affiliations:
  • School of Computer Science, HuaZhong University of Science & Technology, Wuhan, China;School of Computer Science, HuaZhong University of Science & Technology, Wuhan, China;School of Computer Science, HuaZhong University of Science & Technology, Wuhan, China;School of Computer Science, HuaZhong University of Science & Technology, Wuhan, China

  • Venue:
  • IWANN'13 Proceedings of the 12th international conference on Artificial Neural Networks: advances in computational intelligence - Volume Part I
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

In the study of unstructured peer-to-peer networks, due to the lack of network structure, efficient search for resource discovery remains a fundamental challenge. With a combination of the Vector Space Model, this paper presents SMBSRP, aSearch Mechanism Based on interest Similarity, query Relevance and distance Prediction to improve search performance. SMBSRP groups nodes with similar interests together to build the overlay network. By evaluating the query "condition", SMBSRP decides whether to choose part of neighbors to forward query messages. Besides that, this paper proposes a forwarding factor, which is used to estimate which neighbors to be selected to forward query. The forwarding factor is constructed by neighbors interest similarity, connectivity degree and network distance. The experiment results show that, compared with flooding and random walk, without losing the query hit rate, SMBSRP can reduce the redundant information efficiently.