An Improved Attributes Reduction Algorithm Based on Rough Set

Rough set is an efficient information processing tool, and attribute reduction of decision table is a kernel part of research in rough set theory.Based on the analysis of attribute reduct algorithms in reference[7],an improved heuristic reduction algorithm is proposed to effectively and efficiently achieve the better attribute reducts of inconsistent decision table.This algorithm based on the principle of remaining positive region of inconsistent decision table,and only considering objects in positive region,Then combined with the significance of attribute as heuristic information of feature selection.The algorithm is the same with consistent and inconsistent decision tables,but the time complexity does not increase.At last,demonstration section shows that this algorithm is effective.