COMPARISON BETWEEN THE INDUCTION LEARNING ALGORITHM OF FUZZY NUMBER-VALUED DECISION TREE

摘要：

From the value of attributes, evaluation functions, stop-criterion of the leaf node and the matching rules used to test examples, this paper discusses the difference and similarities between the induction learning algorithm 1 and 2 of fuzzy number-valued attribute decision tree.Heuristic algorithm 1 is an algorithm regarding the fuzzy number-valued attribute using the information entropy minimization heuristic; the algorithm gives us a desirable behavior of the information entropy of partitioning and offers a rapid matching speed.Heuristic algorithm 2 is based on the fuzzy information entropy minimization heuristic, this algorithm is used to choose the test attribute and to construct a fuzzy Bi-branches decision tree with comparison extent.By comparing the algorithm 2 closes to the practice from the opinion of making strategy and is effective to deal with fuzzy information.

关键词： Fuzzy number-valued attribute Fuzzy Bi-branches decision tree Comparison extent Degree of truth of fuzzy rules

作者: DONG-MEI HUANG JUN-LI FU TAO XIAO JING ZHOU

作者单位: College of Science, Agriculture University of Hebei, Baoding 071001, P.R.China

会议类型: 国际会议

会议名称: 2007 International Conference on Machine Learning and Cybernetics(IEEE第六届机器学习与控制论国际会议)

会议地点: 香港

会议语种:英文

页码: 1190-1193

在线出版日期: 2007-08-19（万方平台首次上网日期，不代表论文的发表时间）