Study On Semantic Retrieval In Peer-to-Peer Networks

With a rapid development of network technology in recent years, network has now become increasingly ubiquitous. However, the fast-growing network makes traditional C/S mode severely challenged. At the same time, with its exponential growth in network bandwidth and computing capacity, P2P has now received more and more attention in industrial and academic areas and become our research focus. P2P network has many interesting technical functions such as decentralized control, self-organization, adaptation and scalability, which make it a promising choice of harnessing the huge resources of the vast number of internet hosts. However, the data in P2P network are distributed, unstructured and in lack of semantic expression and description, so that they are only shared in the level of coarse granularity (the file level). The low retrieval effectiveness limits the P2P network's potential to work to its best, and decreases the knowledge sharing level among peers.Firstly, a detailed study of P2P network and semantic retrieval is conducted; then P2P semantic retrieval is analyzed emphatically. On that basis, the thesis puts up forward with a semantic retrieval framework model and then makes a detailed analysis of the framework structure and functional modules as well. In addition, the author also gives a vivid description of the search process in this paper. Appling ontology to P2P information retrieval, the model makes use of the semantic relations among ontology concepts in this framework model to direct the retrieval process of the equivalent network, and adopts the functions of the super-peers in P2P network, which enable the users to find their resources more rapidly and precisely, and thus increasing the recall and precision of P2P information retrieval.After that, an in-depth analysis and study of the key issues in P2P semantic retrieval is also demonstrated in this thesis. Besides that, the author also discusses the ontology construction and management in P2P community in great detail, makes a research of the four main factors affecting semantic distance in semantic similarity computation, the asymmetry of semantic similarity, and finally presents a corresponding semantic similarity computation model. In this part, the study of the key issues provides theoretical support for the research and perfection of semantic retrieval in P2P network environment.