摘要
基于 Rough集理论中的可辨别矩阵 ,讨论包含冲突样本的决策表中的决策规则及其可信度和支持度因子 ,提出决策规则的发现算法。算法通过删除决策表的核属性来引入新的冲突 ,以获取具有更高适应度的决策规则。最后用实验结果证明算法的可行性。
In this paper, based on the conception of discernibility matrix, we analyzed the decision rules in decision tables containing conflict samples, and described the confidence and support factors. An algorithm for discovering decision rules has been presented, in which core attributes were deleted to obtain more compatible decision rules. As a result, we have proved the algorithm to be feasible with an experiment.
出处
《系统工程》
CSCD
北大核心
2003年第1期8-11,共4页
Systems Engineering
