| Along with the people to research the technology and application of case based reasoning system gradually, make the case-based reasoning system’s application scope of areas is widening. Therefore, the use of better technology to improve the process of case based reasoning system is particularly important. In the process of case-based reasoning, case representation, case organization structure and case retrieval is the core problem of case-based reasoning system research. Therefore, a better case representation structure, a more reasonable structure, a more efficient and accurate case retrieval scheme to improve the performance of case-based reasoning system to recommend solutions more accurately, so that the case based reasoning system can be better used in more and more fields.This paper after analyzes the characteristics of formal concept analysis and the advantage of formal concept analysis is used in case-based reasoning system, was proposed based on formal concept analysis of case-based reasoning system of technology improvement, the main research work and innovation of this article are as follows:1)Adopt concept lattice said as case representation methods and use the increment algorithm of concept lattice to construction concept lattice and puts forward a new concept lattice incremental algorithm. This article decided to adopt the method of concept lattice said as case representation, it is not only a layered structure and knowledge base can be represented in the form of graph. The incremental algorithm not only solves the problem of reinstitution concept lattice, but also promote the establishment of a dynamic knowledge base. And a new incremental algorithm this paper proposes greatly improve the construction efficiency.2)We propose a similarity model based on the concept of attribute. By combining the characteristics of the process of case retrieval in case based reasoning system, this paper puts forward a based on concept of attribute set similarity model, in this model only by comparing the similarity of description attribute in the concept, not only improve the accuracy of case retrieval process, but also improve the efficiency of the contrast between the concept.3)By comparing with classical algorithms of build concept lattice, similarity model and classification system. Through experiment contrast found using FCA to improve in CBR is feasible, and the improved CBR system will do better in the density smaller data set. |
