The new concept of the fault decision table graph is proposed based on the idea of processing for fault diagnosis
in power system which combines rough set with graph theory. Furthermore, the concepts of fault information coverage
and fault diagnosis rule classification are presented by using the fault decision table graph and its adjacency matrix, and
the method of fast recognition of core attributes of a decision table and the method of attribute reduction are obtained, and
the concept of fault diagnosis rule classification is applied to fault rules extraction. An example is processed by using the
proposed method, and the simulation results show that, the method can effectively reduce the complexity of time and space,
and extract diagnosis rules according to the set threshold value.