Research On Interest-based Driven Search Method In P2P System

Peer-to-Peer technology is to study the hot point in networks at present. How to search the resources in the P2P networks efficiently is the key problem when the P2P networks are developed. The researches of interest-based driven P2P search method on unstructured P2P network is the most important issue. Nowadays, there are tow deficiencies of the interest-based driven P2P search method. (1) The interest of node is solely discovered by these methods from the user or node content's point of view which is unilateral. (2) These methods are lack of context semantic expansibility or have not the ability.Aims at the above problems, paper improve the Social-P2P search method which based on social net theory and small-world theory by introducing the FCA method. This article put forward an improved interest-based P2P search method (IISM). IISM builds the friend list by the matching algorithm of node content and recording the user behavior recently, so that it can build interest domain both in the view of user and node content. The method also overcomes the simplicity of the point of view in interest-based driven P2P search method at present. Then, in view of the successful application of concept lattice theory on the area of information retrieval, the paper introduces the concept lattice theory into the P2P search methods. Then it extract formal context from friend list to build a concept lattice. The concept lattice cluster nodes of similar interest and build the interest domain by concepts which have deflection order. In this method, node forwards message by clustering concepts which cluster nodes of similar interest and have deflection order. And it's capable of extend the context semantic search by this method. So that this method can cut down the search path and reduce the search message. The method can also enhance the search accuracy. Finally, paper verified the method by experiments. To compare with Social-P2P search method and traditional flooding method, the result indicated that IISM has better recall and accuracy and lower network load. So IISM raise the efficiency of P2P search to compare with the Social-P2P and traditional flooding method.