The Application of the New Discretization Method of Fault Diagnosis Based on Rough Sets

There are many uncertainties and knowledge redundancy in Fault Diagnosis. while the rough set theory is a new the mathematical tools to deal with the uncertainty of information. The rough set theory provides a new effective way of dealing with the mechanical fault diagnosis. As the rough set theory can only deal with the decision table after discretization while the actual application is often get continuous attribute values, the most important issue to face this problem is the discrete data processing. In this paper, we propose a new method of attribute discretization methods based on the importance of attributes. This method makes the fault symptom attributes simplification in the case of ensuring the diagnostic precision unchanged, at the same time eliminate the redundant diagnostic determine rules to access to streamline the rules.