Research of an integration model combining Rough Set with extension theory for fault diagnosis

Considering the poor accuracy of extension theory using in fault diagnosis field, a method of combining rough set and extension theory is presented. The method use historical test data firstly to construct the attribute sets and decision-making sets for fault diagnosis, then using rough sets reduce the attribute sets, and calculating the important degree of each condition attribute for decision-making to get its objective weight. Using the reduction information construct the matter-element model, the fault pattern is judged by the weighted relevance degree of the unknown object with each fault pattern in data set of reduction attribution with the relevance function in extension theory. Experiment result indicates that Rough Set based extensional fault diagnose method can effectively determine the best features of parameters describing the fault, distinguish the importance of different parameters, improve the accurate of extension diagnose algorithm, and can play significant performance in practice.