Research On Text Information Retrieval Based On Domain Ontology To Expand Query Requirements

Semantic-based query expansion is the most effective method in enhancing the performance of information retrieval. For this reason, it has become a hotspot in IR field. Based on the research of traditional text information searching method which includes too little semantic information to represent users' requirements and be understood by the computer, and finally leads the retrieval results not to coincidence of user demands as well as significant influence on retrieval performance. This paper presents a text information retrieving method that employs domain ontology to extend users' query requirements aiming at these problems.Extracting domain knowledge from web page and constructing domain ontology. Designing domain ontology's structure in which each node contains concrete concept and actual example. Different level has different composition. With choosing web page as the main data sources of domain knowledge and analysising its constructional feature, the paper proposes a extracing method for domain knowledge aimed at web page and then the domain ontology can be built by using these domain knowledge. First of all,extracting the knowledge structure which is the foundation for exacting the candidate attribute of concept. Structure extraction includes candidate attributes extraction using certain patterns and applying a statistic method to filter out the incorrect attributes and finally define standard attributes. During examples extraction, standard attributes is used as a clue to automatically generate example patterns, and acquire ontology examples taking advantage of the special structure feature of the Web pages. We build core ontology and then extend it using the ontology knowledge extracted from Web page.Raising a query expansion method based on domain ontology. The paper analyses the advantages and limitations of the main query expansion methods and proposes a query expansion method based on domain ontology which includes the analysis of user queries, the semantic vector representation of queries and the match between users' query and domain ontology. At the same time, the appropriate method for semantic extension is given.Applying the method in information retrieval. For the above research, we design and' realize the experimental system. Before retrieving information,we use domain ontology to expand initial query and enhance the semantic of users' queries. By analysizing the experimental result, it shows the effectiveness of this method proposed in the paper.