期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2012
卷号：3
期号：3
页码：3942-3947
出版社：TechScience Publications
摘要：In this paper, an optimize and effective algorithm is proposed for constructing decision tree based on variable precision rough set theory which can deal with inconsistent, uncertain or vague knowledge. FID3 has some drawbacks, it does not provide relaxation to the subset operator .Therefore, we improved FID3 algorithm based on variable precision rough set .This paper proposes a new attribute selection criterion, the enhanced information gain based on degree of .. - dependency and significance of condition attributes on decision attribute is used as a heuristic for selecting the optimal splitting attribute to overcome the drawback of FID3 algorithm. Experiments prove that the improved VPRSFID3 algorithm reduces the complexity of tree and increases classification accuracy of the decision trees as compare to the FID3 algorithm.
关键词：Decision tree; ID3 algorithm; degree of β-;dependency; variable precision rough set (VPRS); enhanced;information gain
