Researching And Realization Of Information Retrieval Based On "Scientific Knowledge Ontology" On P2P

As the rapid development of Internet,there is a growing accustomed to obtain all necessary information from web, but the dramatic increase of amount of information leading to the efficiency of information retrieval downward trend,and serious problems of "information overload" and "information lost" occurred. For scientific research workers, whether they can get the latest research developments of his subject area more quickly and more accurately becomes one of the key factors of which affecting their research work. So we applied the Inner Mongolian Natural Science Foundation project—Researching of Autonomous"Scientific Knowledge Sharing"Based on Ontology. As a major part of the project, this subject- Researching and Realization of Information Retrieval Based on"Scientific Knowledge Ontology"on P2P, it mainly research the ontology reasoning, personal knowledge management, semantic search technology based on ontology, and achieve the basic functions of knowledge-sharing in distributed environment. The main works of this thesis are as follows.First, on basis of studing existing technology and ontology model which has been established, I have realized ontology reasoning machine in .NET environment against this model and demand of this project. It supports the reasoning of multiple relations difined in ontology. The reasoning machine is one of the key technologies of this subject, it is the foundation to realize semantic.Secondly, this paper proposed a human-machine interaction method for documents semantic pretreatment, take scientific research workers as the first step of knowledge filter can effectively improve the value of knowledge base.This method through display concept and attributes definited in ontology to help users complete the document semantic tagging and extraction process, so realizing management function to individual research knowledge.The documents semantic tagging and retrieval is basis of semantics.Third, researching existing information retrieval technology, we design the semantic information retrieval algorithms in two ways. First, for user's search terms, here given a method of semantic query expansion based on ontology using reasoning machine.It not only can improve the recall rate and precision rate, but can also solve the problem of "search deviation" effectively. The second is from organization of resources aspect, we propose a space resources model based on ontology, using semantic vector to build the resource model, so as to enhance efficiency of information retrieval effectively. On this basis, we achieve information retrieval algorithms for different input conditions, such as single, double and more items of condition.Finally, through study and research existing P2P technology, aim at goal of this subject, we designned and developped a prototype system of semantic information retrieval for scientific knowledge on P2P - SK-SIR system. It not only provides personal knowledge management functions, but also realizes sharing of knowledge in distributed environment.