Extraction method of decision rules for fault diagnosis from incomplete data based on rough set

Extracting rules from incomplete data are usually more difficult than extracting rules from complete data in mechanical fault diagnosis. In order to extract simple and effective diagnostic rules from incomplete data, a method to directly extract decision rules for fault diagnosis from incomplete data based on rough set is proposed in this paper. The method realizes an object-oriented reduction approach in an incomplete fault diagnosis decision table using the defined object-oriented discernibility function, and then simple and understandable decision rules are extracted directly from an incomplete fault diagnosis decision table using the obtained all object-oriented reductions. The application of the proposed method is demonstrated by a mechanical fault diagnosis example with incomplete information.