Joint Decision Matrix Computation for Rules Extraction in Two-Classes Decision Information System

Matrix computation for rules extraction of rough set algorithm's inefficiency restricts the application of rough set theory, so the search for efficient algorithms is of great practical significance. Therefore, the joint decision matrix is defined and a new algorithm JDMCRE for rule extraction is presented on the basis of joint decision matrix for two-classes decision information system. By combining both the conditional attribute equivalent matrix and the decisional attribute ones into one matrix, it can considerably reduce the number of comparison in the process of matrix forming, which can effectively increase the efficiency of algorithm. The complexity analysis and experiment results show that the algorithm is much better than existing matrix algorithms.