Research of the Case Retrieval Model Based on CBR

Taking the direct impact of Case retrieval model on Case-Based Reasoning(CBR) into consideration. And the appropriate selection of membership functions has a direct impact on the preciseness of similar degree, so as to influences the accuracy of the multi-dimension similar degree, which will ultimately affects the retrieval results of the new problem´s most analogous instances. This paper proposes a novel Case retrieval model that integrates the three approaches of multi-concept-learning decision tree algorithm, multi-dimension weighted similar degree and fuzzy synthesis evaluation. It successively advances certain membership functions on the basis of the attributes of cases. The retrieval model proposed in this paper integrates the decision three approaches of multi-concept-learning decision tree algorithm, multi-dimension weighted similar degree and fuzzy synthesis evaluation, which, in essence, is a mergence of inductive method and adjacent method. The advantages of these two methods are fully utilized to promote the retrieval effectiveness. In the actual development process, the retrieval model is similar to the way of human thinking, and is easy to explain. It is a good tool to solve any decision problems of unfavorable structures. Finally the paper recommends, alongside the advantages and drawbacks, the future ameliorative and innovative direction of the above-mentioned model.