Research On The Semantic Transitivity Reasoning Model In Objective Knowledge System

The research on relatedness in objective knowledge system includes two aspects:one is to judge whether concepts are related or not, the other is to reason relation type between concepts. Up till now only few attempts on the latter research aspects have been made and similar researches ignored the four methods over time including syllogistic reasoning, plausible inference based on relation elements, reasoning based on semantic description languages or description logic and reasoning based on semantic transitivity model. However, two problems exists:one is that previous research based on qualitative analysis is lack of objectivity. The other is that there is a contradiction between the richness of semantic description language and the incomplete reasoning algorithms for current researches which are almost from the view of semantic reasoning.Aiming at the problems above, from a new view, the path of the semantic transitivity, semantic transitivity reasoning model in objective knowledge system has been studied in this paper. First of all, based on the analysis of semantic structure of objective knowledge system, semantic transitivity reasoning model has been constructed. Secondly, the method on constructing system of relation element and generating rules of semantic transitivity has been proposed and then the feasibility of the method was verified by applying it in WordNet. Finally the program of semantic transitivity reasoning has been designed and realized by applying the rules of semantic transitivity above into the improved inference mechanism. The main function of the program above is to judge the relation type between any two concepts.Based on semantic relation, the semantic transitivity reasoning model in objective knowledge system has been proposed in this paper. One one hand, the model has been formed for the first time, which enriched the research on semantic transitivity. On the other hand, it increases the accuracy of knowledge retrieval. So this paper has important application value in the field of knowledge organization and information retrieval, and meanwhile it expands correlation research between concepts.