Investigation of Knowledge Inclusion Relations in Faults Knowledge Discovery

By means of the essential notions in rough set theory, to partition a special universe U into the different knowledge particles sets is investigated. A novel granularity calculation model is built out. It makes use of the different division rules to partition the same U. The relations reflected in the different knowledge particle sets show as important information. It is that knowledge inclusion relations come forth in knowledge discovery naturally. And the correlative researches with regard to knowledge discovery must aim at the idiographic problem and make use of the special data set to progress. If to look forward to extract out the knowledge rules from a data set with the experts´ level, the experts´ knowledge should be save into original data beforehand and the saved manner should be scientific. It is that the data to describe the experts´ knowledge must satisfy the demands of knowledge discovery tool. The characteristics of rub-impact faults taken place on a rotor show the validity of investigation.