一种基于遗传算法的兴趣规则挖掘算法

Discovering Interesting Knowledge with a Genetic Algorithm

数据挖掘是在数据中发现隐藏的结构和模式。但发现的许多模式对用户来说可能是已知的,从而使这些模式毫无意义,毫无兴趣性。文献中多强调分类规则的准确性和可理解性,但发现兴趣规则在数据挖掘算法中依然是一个令人生畏的挑战。本文采用一种遗传数据挖掘方法,在分类规则产生的同时对其兴趣性进行度量,直接产生兴趣规则。实验表明该方法是可行的、高效的。

Data - mining is the process of discovering hidden structure or patterns in data. However the discovered patterns may be obvious to the user and then become unvalued, namely no interestingness . Although the literature emphasizes predictive accuracy and comprehensibility, the discovery of interesting knowledge remains a formidable challenge for data mining algorithms. In this paper, the interesting rules are directly generated using a genetic algorithm. The scheme is proved be practicable and efficient.