标题：A granularity attribute reduction method based on binary discernibility matrix
作者：Chang, Tong ;Shifei, Ding ;Hong, Zhu ;Hongjie, Jia ;Jian, Wang
作者机构：[Chang, Tong ;Shifei, Ding ;Hong, Zhu ;Hongjie, Jia ] School of Computer Science and Technology, China University of Mining and Technology, Xuzhou 221 更多
来源：International Journal of Advancements in Computing Technology
摘要：At present, most of the attribute reduction algorithms based on granularity are simply computing the granularity of knowledge. Repeated calculation will increase the time complexity. Binary discernibility matrix is used to express binary form, which has a clear ascension whether in space or in time than the traditional discernibility matrix efficiency. On the basis of granularity-based attribute reduction, a method is proposed to preprocess the dataset by using binary discernibility matrix. Firstly, find the core attribute and a minimal reduction. Then use the granularity thought to calculate each particle of the importance of attributes. Most important is joined to the reduction set, thereby achieving the attributes reduction. The example analysis shows that the method can improve the performance of the traditional attribute reduction algorithms effectively. It is a feasible approach to reduce attributes.