Research And Application Of Intelligent Search Engine Based On Ontology

With the development of Internet, web information resources are also increasing. Confronted with the larger knowledge which content is redundant and structure is complex, how to acquire the information accurately, timely and comprehensively they need has become the focus of attention. Traditional search technology based on simple keyword matching can not meet growing customer demand. Ontology is a describing model on semantic and knowledge levels,which has perfect conceptual hierarchical structure and support of logical reasoning.Semantic search technology based on ontology has become a hot spot of foreign and domestic scholars'research.Fistly, the progress of semantic retrieval at home and abroad was discussed, and the principle, classification, evaluation criteria and shortages of traditional search engines were analyzed. Secondly, on the basis of research of ontology relevant fields and domain ontology construction strategy, household appliance domain ontology was constructed and stored. Furthermore, after researching semantic annotation technology which restricting development of semantic retrieval, and analyzing current semantic annotation methods and tools, a semantic annotation tool named AnnTool was developed and realized. Finally, with the household appliance ontology and semantic annotation technology introduced into information retrieval, a search engine framework based on domain ontology was proposed. This framework includes five aspects: information gathering, preprocessing (semantic annotation),semantic indexing, query extending , results re-sorting. After analyzing current crawling strategyes ,an integrative crawling strategy was proposed and an value formula of document was gived at information gathering module. the preprocessing module realized semantic annotating using AnnTool. Meta data indexing was realized at semantic indexing module. Query conditions were extended according to domain ontology concept model.The results was sorted by scoring mechanism of Lucene using ontology weight of matching similarity and integrated value of documents as parameters . By analyzing and evaluating experiment results of system ,we could find that search engineer based on ontology provided more recall ratio and precision ratio than traditional search engines.