摘要
提出了一种基于粗糙集理论的新决策树剪枝方法.在剪枝的过程中,不仅考虑了树的分类精度,而且还考虑了生成树的深度对剪枝的影响;最后针对具体的数据集对新方法进行了验证,得到了较好的效果.
Pruning decision tree is an effective way to avoid the phenomena of overfitting. This paper gives a new decision tree pruning method based on Rough Set Theory（RST）, which holds that the complexity of the decision tree should also be regarded besides the classification accuracy of the tree in the process of pruning decision tree. It takes into account not only the classification accuracy of the tree but also the depth of the tree. Finally, a data set is given to validate the new method which has shown good performance.
出处
《东北师大学报（自然科学版）》
CAS
CSCD
北大核心
2005年第3期28-32,共5页
Journal of Northeast Normal University(Natural Science Edition)
基金
吉林省科技发展计划项目(20040529)
东北师范大学青年基金资助项目(111420000)
关键词
过匹配
剪枝
深度拟合率
错误率
overfitting
pruning
depth - fitting ratio
error ratio
