An Improved Attribute Reduction Algorithm Based on Attribute Importance Function

A kind of attribute reduction algorithm of Jiahua Lu was firstly introduced and its performance was analyzed briefly, then aimed to its improvable points, a new attribute reduction algorithm based on the binary discernibility matrix and an improved attribute importance function that utilized the potential information of the information system, which considered attributespsila occurrence number and the lengths of the matrix elements, was proposed in the article. The new algorithm can show the real significance of the attributes, and the matrix scale is significantly reduced, its feasibility and validity and its lower time complexity than the similar ones are proved by the comparative analysis.