Research Of Semantic Ontology-based Food Vertical Search

With the rapid development of the Internet, online content is greatly enriched. So users tendto use search engines to retrieve information in order to quickly obtain valuable informationresources in the mass. The diversification of the market demand prompted professional searchengine and domain-specific portals increasingly popular in recent years, which is also known asvertical engine and vertical portal.Traditional vertical search engines still use keywords or text content-based retrieval methods,which can not fully express the semantic information, but only be expressed from keywordsemantic level up to handle user requests. The main reason exerts this problem is that documentsare lack of semantic annotation. Therefore, search engine could not do semantic analysis onuser’s query. Combine ontology with search engine, becomes an important developingorientation of search engine.As web information increases dramatically every day, search engines need a fast indexingapproach to improve the efficiency of real-time search. Moreover, the large amount of data andcomputation, it costs a lot time for search engines to index the large large-scale text, whichreduces the searching performance. So how to achieve high performance the massive text dataindex is indispensable to search application issues.Therefore, this dissertation focus on how to combine ontology with search engine. Then,introduce a framework of food vertical search engine framework. In addition, an acceleratedindex constructed model is proposed. The main work of this dissertation are as follows:(1) Bulid a food ontology by analyzing the food information, which defines the concepts ofclasses, relationships, and attributes in order to better integrate knowledge in the field of food.(2) According to the characteristics of the food websites use rule-based informationextraction technology to extract the ontology individuals automatically, which lay the foundationfor semantic analysis and semantic query expansion.(3) Propose CPU+GPU collaborative model by analyzing the index build mechanism inorder to give full play to the CPU and the GPU’s computational advantage, to improve theperformance of the index building.(4) Explore the the semantic query expansion process under the guidance of the foodontology.