摘要
决策树归纳算法 ID3是实例学习中具有代表性的学习方法 .文中针对 ID3易偏向于值数较多属性的缺陷 ,提出一种新的基于属性 -值对的决策树归纳算法 AVPI,它所产生的决策树大小及测试速度均优于 ID3.该算法应用于色彩匹配系统 。
The decision tree induction algorithm ID 3 is representative among the learning from examples algorithms, but it is liable to select the attributes with more values. So a new decision tree induction algorithm with attribute value pairs (AVPI) as branch criterion is presented in this paper. The decision tree size and testing speed about AVPI are better than ID 3. Tests show AVPI is effective when it is applied in the color matching system.
出处
《小型微型计算机系统》
CSCD
北大核心
2001年第4期459-461,共3页
Journal of Chinese Computer Systems
基金
哈尔滨工业大学校基金
关键词
属性-值对
决策树归纳算法
符号学习
实例学习
Learning from examples
Decision tree
Information entropy
Attribute value pairs
Color matching method
