The Decision Table Compression Based On The Rough Set Theory

In this paper we propose a method of local reduction of a decision table, give a fault diagnosis problem and get critical fault features affecting certain equipment troubles through the application of the local reduction technology. We define the concepts of homomorphism mappings and endomorphism mappings on some sets and investigate their main properties. We propose the condition of decision table compression keeping attribute reduction and local reduction invariant on approximation space. An example is employed to illustrate decision table compression keeping reduction invariant. Finally, we define consistent fuzzy decision system and its relative reducts, we also develop a method by discernibility matrix to compute all the relative reducts.