论文部分内容阅读
扩展了Rough集正区域和边界的定义,在得到信息系统最大正区域的前提下,给出了认知正区域、认知属性核和认知属性约简的定义,并给出了从经典属性约简到认知属性约简转换的高效算法。此外,在认知正区域的定义下,由于决策表的不相容性,在变精度模型下实现属性约简的增量处理是相当困难的,结合提出的高效算法,解决了这一问题。最后,仿真实验说明了算法的有效性。
The definition of positive region and boundary of Rough set is extended. The definition of positive region of cognition, cognitive attribute kernel and cognitive attribute reduction is given under the premise of obtaining the largest positive region of information system. Reduction to the efficient algorithm of conversion of cognitive attribute reduction. In addition, due to the incompatibility of decision tables, it is very difficult to realize the incremental processing of attribute reduction under the variable precision model. This problem is solved by the proposed efficient algorithm. Finally, the simulation experiment shows the effectiveness of the algorithm.