Study on attribute reduction based on mapping matrix

The flaw of attribute reduction based on the discrimination matrix, decision matrix or heuristic reduction lies in failing to obtain optimum reduction set or deal with incompatible decision tables. Having applied the universal law of general character and individuality of things to ordinary decision table, we put forward attribute reduction based on mapping matrix to directly find reduction set. According to the mapping relation between an attribute value and a decision value from decision table preprocessed, the mapping matrix was obtained. After singular value decomposition (SVD) was carried out to the mapping matrix, the variety value vector of attributes corresponding to decision values was extracted to ascertain origin core and main core attribute, which compose the optimum reduction set. The comparative analysis of examples confirms the higher reduction effect and adaptability of the algorithm advanced.