Fast Matrix Computation Algorithm for Rules Extraction in Incomplete Decision Systems

Some new concepts were defined such as indiscernibility relation of general decision,the indiscernibility classes,joint decision tolerance matrix and a fast computation algorithm was proposed for reduction and rule extraction based on joint decision matrix in incomplete decision systems.By putting the tolerance matrix of conditional attribute and assignment decision matrix of decision attribute into one matrix,the comparison times was considerably reduced in matrix forming process,which could considerably improve the algorithm efficiency.Through an example,the algorithm carrying-out was accounted for.The complexity analysis and a corresponding comparative experiment show that the method is quite effect.