Study And Implementation Of Tourism Information Retrieval System Based On Domain Ontology

This thesis not only analyzed the present situation of application and the technologies of information retrieval (IR), but also deeply researched on the ontology-oriented IR technologies. For the low efficiency of IR problem in tourism field, in this thesis, we studied and implemented an IR system based on Beijing tourism information domain ontology. Research and innovation are as follows:(1) Analyzed the requirements of the purpose, the application range and core concepts of the domain ontology, and studied the process of determination of classes, relationships and attributes. Then used Protege and "seven-step method" to build the Beijing tourism information domain ontology.(2) An Semantic Distance based Relationship Path Semantic Similarity Calculation Model (SDRPCM) is proposed and built. Three impact factors of SDRPCM model is defined and calculated. Proposed the concept of relationship path weight, and its parameters and formulas are given. The experimental results indicate that SDRPCM has better performance than classical models such as DBCM and ICBCM, meaning that, the semantic similarity of SDRPCM is more consistent with the expert experience in the field.(3) Implemented keywords query expansion based on SDRPCM, and optimized the semantic sorting algorithm in query expansion process. An improved Lucence-based Semantic Sorting Algorithm (ILSS) is proposed. The experimental results show that, ILSS algorithm performs better than TF-IDF algorithm in semantic sorting, meanwhile, the query expansion based on SDRPCM model and ILSS algorithm is significantly better than traditional keywords search. An Entity-based Inversed Index Structure of Path (EIISP) for the set of retrieved documents is built, that effectively reduces the query time of searching compound keywords, and improves the query expansion efficiency.(4) A Domain Ontology based Tourism Information Retrieval System (DOTIRS) is implemented. The system has the query expansion and semantic reasoning capabilities on the training documents which has the inverted index structure of path. The DOTIRS system has been applied in practice.Domain ontology based tourism IR system could not only give the standardized description of the domain knowledge, but solve the semantic heterogeneity problems in information sharing process. Through the logical description and semantic reasoning of the set of the abstracted concepts and relationships, domain knowledge could be effectively expressed in semantic level. Therefore, this thesis laid a theoretical foundation for the further optimization of the IR technologies.