ONTOLOGY-BASED DYNAMIC AND SEMANTIC SIMILARITY CALCULATION METHOD FOR CASE-BASED REASONING

摘要

In this paper, we propose an Ontology-based DYnamic and Semantic similaritY computation method for CBR (ODYSEY). ODYSEY is developed for the computation of dynamic similarity as well as semantic using dynamically changed ontology structure according to occurred context using CBR. Context is defined as any information that can be used to characterize the occurrence situation of a new problem and cases. To compute dynamic and semantic similarity, ontology is restructured by the context. The domain ontology is developed by consideration of contexts, and partial ontologies are extracted from the ontology including a set of shared contexts with same features and values between a new problem and cases. We implemented a mobile e-mentoring system, named MintStory~© to show the applicability and feasibility of ODYSEY. As experimental results show, the satisfaction value of similarity of MintStory~© is higher than the coordinator-recommended result. It shows that the ODYSEY is significantly efficient in a similarity calculation problem in CBR.