THE RESEARCH AND REALIZATION OF DECISION TREES BASED ON CREDIBILITY

摘要：

In the inductive learning, if example bank has noise, it is hard to obtain decision trees with high precision, that is, it is hard to obtain knowledge with high credibility. So, this paper puts forward C-ID3 algorithm, by which we compute confidence level of example bank firstly, and further get a kind of decision tree on the basis of traditional IDS and the credibility obtained. All the theory analysis and simulation indicate that this method posses interesting feature of strong operability, and it also can improve the reliability of knowledge obtained.

关键词： Data mining Inductive learning Decision tree ID3 C-ID3 Noise Credibility Modes purity

作者: FA-CHAO LI PING LI

作者单位: School of Economics and Management, Hebei University of Science and Technology, Shijiazhuang 050018, School of Science, Hebei University of Science and Technology, Shijiazhuang 050018, China

会议类型: 国际会议

会议名称: 2008 International Conference on Machine Learning and Cybernetics(2008机器学习与控制论国际会议)

会议地点: 昆明

会议语种:英文

页码: 345-349

在线出版日期: 2008-07-12（万方平台首次上网日期，不代表论文的发表时间）