Subway fault diagnosis method based on data mining

The invention discloses a subway fault diagnosis method based on data mining. The method comprises the steps that history fault data of a subway is disposed to form an original decision table; the original decision table is pre-processed by a rough set, redundant attributes are reduced to obtain a new attribute set, and the original decision table is transformed to a new decision table with a boolean attribute; association rules between an attribute set of the new decision table and faults are mined to store in a rule base; state data and the rules in the rule base are compared in real time to make fault diagnosis. According to the subway fault diagnosis method based on data mining, the rough set is adopted to perform data pre-processing, data mining is used in fault diagnosis, the faults occurring on the subway are judged, and not only difficulties brought by the fact that subway working personnel judges the faults according to experiences are solved, but also the speed and accuracy of fault solving are greatly improved.