一种基于扩展差别矩阵的规则获取方法

A Rule Acquisition Method Based on Extended Discernibility Matrix

本文对Skowron差别矩阵[1]进行扩展,提出了一种不求决策表属性约简,直接获取决策规则的方法。此外,该方法只需根据同一决策类中扩展差别矩阵元素来提取规则,且不会造成规则冲突,适合于分布规则提取。

In this paper, by expanding Skowron＇s discernibility matrix, we put forward a method to extract rules from a decision table The method processes the data in advance so that the extraction of classification rules can be efficiently done in a distributed way. Moreover, the method allows direct acquisition of rules without computing attribute reductions