Study On Information Retrieval Model For Centralized Mixed P2P Networks Based On Field Partition And Semantics

The technology of Peer-to-Peer is a hotspot of research field on international Computer Networking, and the main aim of the technology is to make the best use of the potential resources in the network. The Retrieving Technology of P2P is also an important part of researching P2P. With the explosively various information on the Web, the pivotal point of P2P is how to improve the use ratio of resources in the P2P and how to ensure the Recall Factor and Pertinence Factor of retrieving , because these factors influence the validity and development of P2P .In allusion to the low efficiency of retrieving information in P2P and on the basis of analyzing the characteristic of P2P, we propose a information retrieving model for Centralized Mixed P2P Networks Based On Field Partition And Semantics. The main works and contributions of this paper are including:Firstly: The model adopts Centralized Mixed P2P structure, and this structure is much better than other frames on performances and the capacity of peer and flexibility.Secondly: We put forward the concept of Field Partition and the method of partition. All the peers in the P2P are divided into different parts by field, and peers in the same field are linked up by semantics. This method can limit the retrieving scope in the special field. Then the retrieving route and time will be obviously shortened.Thirdly: we induct Ontology and Semantic Web to P2P, thus the information in the network are semantic and can be understood and processed by computer. Anyway it can be convenient for cooperation between people and computers. Recall Factor and Pertinence Factor will be improved simultaneously. Ultimately the results are more precise and more usable.