一种近似决策规则生成算法及其应用

Generating algorithm of approximate decision rules and its applications

实际应用中,Rough集理论所推导出的基本决策规则在对新对象进行匹配时,往往会遇到新对象不能被匹配的情况,使得不能对新对象作出有效的决策.为了提高规则对新对象的匹配能力,提出了一个更有效的近似决策规则生成算法,并将其应用于基于Rough集理论的垃圾邮件过滤模型中.实验结果表明,该算法所产生的近似决策规则能够有效地提高规则对新邮件的匹配能力,以及对邮件的过滤性能,因此该算法是可行的.

In the practical applications, some basic decision rules induced by rough set theory can＇t match a new object, so it often can＇ t make an effective decision to the object. In order to improve the matching ability of decision rules, an efficient algorithm generating more approximate decision rules is proposed in this article, which is applied to the field of filtering spam based on rough set theory. The results in emulational experimentation show that those approximate decision rules obtained by the algorithm can enhance the matching ability, so the algorithm proposed in the paper is effective and feasible.