The Research On Measurements Of Semantic Relatedness Of Topic Map

The complication of knowledge organization increases gradually, with the explosion of information, the diversity of information type and storage, the application request enhancement of knowledge integration, navigation and retrieve. Traditional methods of knowledge organization, such as index, glossary and thesauri, reveal the deficiency on semantic expression, so the research and application on methods of concepts-based association, which can reveal semantic structure of information resource, becomes a focus. Topic map, as a main method of knowledge organization of concept-based association, is research hotspots.The research on technical system of topic map makes mature development and results of its applications in knowledge organization, knowledge retrieval and the construction of concept knowledge base are remarkable, but there are some limitations existing in it. Technical framework of topic map can define various associations, but it doesn't provide the quantization of relatedness in topic map and the associations between topics only confined to direct association, which leads to the neglect of research on semantic transitivity between topics association indirectly.In the light of problems mentioned above, based on national social science fund, the evaluation methods on semantic relatedness of topic map between topics and between topic and resource are the main researching contents in this paper. The construction of the evaluation methods relies on analysis of topic map semantics. In view of these features menthioned above, first of all the characteristics of semantic elements contained in topic map and the semantic structure are studied; secondly, there is systematic study of the semantic transitivity, including the definition, influencing factors, judging process of semantic transitivity and transitive principles, making computer simulation of semantic transitivity possible; then, the evaluation method on semantic relatedness between topics based on semantic transitivity is put forward according to research results done above, and effect test and improvement of the method are done based on HowNet. After that, the method of semantic relatedness between a topic and information resources is constructed based on semantic relatedness between topics and some application features of topic map in knowledge retrieval and the strength of it is pointed out comparing with other methods existing; finally, the application research of topic map knowledge organization based on association measurement is conducted. Through constructing topic map of typhoon as the knowledge base, typhoon knowledge navigation system is designed and implemented embedded the evaluation methods on semantic relatedness of topic map, which can improve the retrieval efficiency of navigation system.The research work of this paper enriches the semantic studies of topic map and introduces association measurement to topic map. All of these can improve the technical system of topic map as a typical kind of concept-association knowledge organization method and make a clearer understanding of semantic structure of topic map and associated resource entities, further more it can be applied to knowledge retrieval area and improve the retrieval accuracy, which is valuable to knowledge organization and knowledge retrieval; at the same time, the work on semantic transitivity and evaluation method based on the transitivity also extend the research on semantic relatedness and transitivity area.