Used In Intelligent Web Retrieval System To Expand Students' Knowledge

This thesis mainly studies the basic principle and structure of the intelligent Web Information Retrieval (IR), which puts forward the new method of intelligence and ability, high efficiency and particularity. We stresses on "Original expression", "difference of expression" and "latent retrieval", which influence the efficiency of IR a lot. Though many approaches have been raised, each has its disadvantages. On the basis of these approaches, we discussed the advantages and feasibilities of IR based on dynamic knowledge base and added natural language processing in user interface. Following this idea we developed Web IR facing students to expand their knowledge.In the matter of techniques, we adopted concept semantic network to create knowledge base, used local context analysis to extract characteristic words from the documents that users inquired, then apply subsumption approach and resemble approach in discovering relationships between terms. The dynamic knowledge base is updated using these extracted terms and term relationships, as well as user's feedback . IR system would expand user's query according to the dynamic knowledge base. Natural language processing used Automatic Segment based on Two-letters-word-family structure, which is succinct, high speed and easy to achieve.This Web IR has three tiers: client, application server and data server. Client mainly accomplishes the communication of user and system, as well as natural language processing; Application server concludes six function modules(query expanding server, personal managing server, retrieving server, results handling server, documents handling server and controlling server ), dealing with works respectively,' such as documents pre-handling, query expansion, personal managing information, retrieving and updating knowledge base; Data server concludes ontology knowledge base, documents base and personal data base.