基于决策支持度的决策树生成算法

Decision trees generation algorithm based on decision support degree

从条件属性对决策支持程度不同的角度出发,引入了决策支持度的概念,提出了一种以其为启发式信息的决策树生成算法。实验分析表明,相对于传统的决策树生成算法,此算法改善了决策树的结构,有效提高了决策分类的精度。

Based on the viewpoint that conditions attributes have different decision support ability,the concept of decision support degree is introduced,and a novel algorithm for building decision tree is proposed, in which the decision support degree is regarded as heuristic information.The experimental analyzes on several public date sets show that unlike the decision tree built by traditional algorithms,the decision tree built by this algorithm has much better tree structure and classification precision.