关键词: 粗糙集, 属性约简, 分类质量, 属性重要性, 决策表

Abstract: This paper firstly introduces some kinds of algorithms for attribute reduction briefly, analyses their advantages and disadvantages. According to this, a new algorithm for attribute reduction based on classification quality is presented, which can be used in complete information systems and incomplete information systems, so as to obtain the satisfying attribute reduction. Finally, the algorithm is demonstrated to be true and effective with specific examples.

Key words: rough set, attribute reduction, classification quality, attribute significance, decision table