Rough Set And Grey-relational Theory In Case-based Reasoning

Case-Based Reasoning is an important analogical reasoning technique in artificial intelligence. It is widely used in the domains that have rich experience (cases), but lack of comprehensive and sound theories. However, there are still some difficulties in solving practical problems only with case-based reasoning. The deepening and improvement of study of some theoretical methods, such as rough set and grey-relational analysis, provide a good solution for the further application of case-based reasoning.This dissertation begins with a brief introduction of the workflow of case-based reasoning, the conceptions and theories of rough set and grey relational analysis. Then integrates the three theories, and design a model of case retrieval based on rough set attribute reduction and grey similarity calculation. Firstly, this model scatters the continuous attributes with rough set theory, calculates the different importance of the condition attributes to the decision-making attributes, and then deletes the condition attributes which have small importance, calculate the weight coefficient of the reserved condition attributes. Secondly, this dissertation combines grey-relational theory and Euclidean distance, and proposes a new method for computing degree of similarity in case-based reasoning, and calculates the similarity between different cases. Attribute reduction and the improvement of similarity algorithm improve the search efficiency and similarity resolution. Thirdly, this dissertation cites the case of weather forecast as an application example to explain the process of the model.At last, the main conclusion is drawn, and the next necessary work is pointed out.