论文部分内容阅读
信息观下研究邻域决策系统的属性约简是一种新颖的思路.通过分析论域下某样本邻域中其他样本与该样本决策属性值的异同,定义不一致邻域矩阵.在计算属性重要度时,利用不一致邻域减少在原条件属性基础上增加一个属性后条件熵的计算时间.分析得到邻域系统下条件熵与正域的关系,提出一种信息观下基于不一致邻域矩阵的属性约简算法,并分析该算法与其他算法的内在联系.实验结果验证了所提出算法的有效性.
Under the view of information, attribute reduction of neighborhood decision-making system is a new way of thinking.This dissertation defines the inconsistent neighborhood matrix by analyzing the similarities and differences between the other samples in the neighborhood of a sample domain and the attribute value of decision-making in the sample neighborhood. , We use the inconsistent neighborhood to reduce the computing time of the conditional entropy after adding an attribute based on the original conditional attributes.We get the relationship between the conditional entropy and the positive domain in the neighborhood system and propose a new attribute based on the inconsistent neighborhood matrix Reduce the algorithm, and analyze the internal relations between the algorithm and other algorithms.The experimental results verify the effectiveness of the proposed algorithm.