An Algorithm of Attribute Reduction Based on Rough Sets

The concept and the nature of attribute reduction from the view of logic was analyzed, and based on this analogical matrix was proposed. Then proposed a new algorithm of attribute reduction using the analogical matrix, and the correctness and feasibility of it was proved. The algorithm can reduce time complexity and spatial complexity of attribute reduction, and do not break the coherence of information contained in decision table. The analysis of the realistic example shows that the algorithm is effective and feasible.