基于信息熵和偏差的加权概念格内涵权值获取

为有效反映数据本身隐含的客观信息,快速提取用户需求的具有一定偏差程度的重要知识,提出了一种基于信息熵和偏差分析的加权概念格的内涵权值获取方法.在缺乏先验知识时,由数据集中属性特征的信息熵来刻画加权概念格的单属性内涵权值,采用均值计算多属性内涵权值,并用标准差计算多属性内涵重要性偏差值;由用户设立加权概念格内涵的重要性阈值和内涵重要性偏差阈值,构造出一种强加权概念格.通过实例描述了该方法可有效指导正确决策,进一步拓广了概念格的理论与应用.

Intension Weight Value Acquisition of Weighted Concept Lattice Based on Information Entropy and Deviance

To effectively reflect the objective information hidden in the data and sensitively extract important deviant knowledge concerned with the user, this paper presents a method of intension value acquisition of weighted concept lattice based on entropy and deviance analysis. On occasion of the absence of prior knowledge, the single attribute intent weight value is decided by the entropy of attribute characteristic in data set. The multiple attribute intent weight value is computed by mean value and the intent importance deviance value among the multiple attribute is computed by the standard deviation. With the given intent important threshold and intent important deviance threshold, that are decided by user, a strong weighted concept lattice is proposed. In the end, an example shows that the presented method is feasible and effective for decision making and it further expands the theory and application of concept lattice.