Research And Application Of Semantic Retrieval On The Web Based On Ontology

Due to neglect of semantic information contained in the resource,the traditional key-word-based web retrieval system can only get a lower rate of recall and precision. And on-tology,which come from artificial intelligence and knowledge engineering, can well processthe problem of natural language understanding and reason with semantic-based mechanisms,so as a good tool to improve the traditional way for information retrieval, Semantic reasoningmechanism is the core of semantic retrieval technology. Now, the semantic reasoning researchand applications emerging, more prominent is the research and application of expert systems.Expert system begins effective reasoning with stored knowledge according to the rules ofthe knowledge base, the core parts of the system are knowledge base and inference engine. Asmost of the expert system’s knowledge base and inference engine is designed for specific useareas and systems, it is difficult with other expert systems in related fields to achieve know-ledge sharing and reuse, and difficult to achieve the separation of reason and fact. Ontologicalrepresentation of knowledge that has strong capacity, as a basis for building expert systemsknowledge base, can provide a better basis for reasoning.This paper had the research and analysis with the existing expert system mechanisms andtraditional information retrieval technologies, it build an ontology-based expert system model，and web semantic retrieval framework is proposed based on expert system, in which expertsystem as the semantic reasoning part of the framework, which carrying the system’s coretechnology. This semantic retrieval framework actually make a pre-built platform on surfacelayer of traditional information retrieval, it will make the query string through the Chinesesegment, the field division, sentence type determine, sentence matching and expert systemtechnology for processing, finally it to get the semantic query then enter the next step.Finally, the paper constructed blog semantic retrieval system, which to verify the feasi-bility of the research of web semantic framework. The system uses spring mvc framework,had jsp as the main page to show results. User requests information should through the Dis-patcherServlet for control navigation, the background support of system is to use jena as ex-pert system inference mechanisms, lucene as a traditional information retrieval framework.Knowledge base of expert system which uses ontology language to make model, the rulesshould be combined costom rules and jena system rules. Finally, it proved that the recall andprecision rate of system which compared to traditional web search system has significantlyimproved.