Research On Framework Of Ontology Based Case-based Reasoning System

It is a hotspot to study the case-based reasoning technology, which make up for the shortage of expert system and supports the whole life circle of knowledge management in the artificial intelligence field. However, in the tranditional CBR system, the variable schema and various management strategy of domain knowledge lead to an isolated island of information, and the knowledge sharing, exchanging and integrating are not so effective. The unscalable case expression and the unreusable reasoning process lead to the development and the maintenance of the systems difficult and costly. Aiming at the goal of helping the developers build the highly scalable and reusable CBR systems, the ontology technology is introduced in our research. This paper mainly researches on the ontology-based CBR system framework.Firstly, the current CBR domain is summarized, the importance and inevitable trend of combining the ontology with the CBR is analyzed, the theories and the features of the two technologies are explained, and then the framework of ontology-based CBR system is proposed in this paper.Secondly, the old case representation models are researched, and the ontology-based case representation model is proposed, which is domain independent and case structure independent. The flow of case knowledge access is analyzed, and is also designed and implemented in our framework. The theory of task structure is studied, and then the CBR task structure is proposed. The concept model of task-method ontology is designed, and it is implemented in OWL to support the reusable task and method in CBR process. In the end, the reusable CBR reasoning is implemented.Lastly, with the help of the expert of the weather domain, the ontology of disaster weather domain is built and the disaster weather CBR system called the WPCBRS has been completed quickly, which proves that the correctness of our OCBRSF.