基于类别特征矩阵的决策规则提取算法

Extracting Rules Algorithm Based on Class Feature Matrix

在粗糙集中决策表属性最小约简与规则提取是NP hard的。充分考虑了可辨识矩阵的特性,提出一种基于类别特征矩阵的决策规则提取算法。首先对类别特征矩阵进行了不相容对象的剔除和核属性的提取,然后利用核属性集合对类别特征矩阵来提取决策规则。实例结果表明,所提出的算法获得的规则更为简洁、高效。

The minimal reduct and extracting rules of decision tables has been proved to be NP - hard in rough set theory. An extracting rules algorithm based on class feature matrix is presented, which has made use of the characteristic of discernibility matrix, It first delets the inconsistent objects and extracts core attributes from class matrix, then adopts core attributes to extract decision rules for class feature matrixes. The experiment shows that the algorithm provides more precise and simple rules.