摘要
基于属性值并的权熵思想,通过构建模型,给出了一个属性的某几个属性值并的权熵之和不小于该属性单个属性值的权熵之和的结论,从理论上证明了ID3算法的合理性,为ID3算法提供了理论基础.
Based on weighted entropy of the union of the attribute value, this paper gives the conclusion that the weighted entropy of the union of several attribute values is not less than the sum of the weighted entropy of the single attribute value. So, the paper provides a theoretic basis for the ID3 algorithm that is an optimal method in decision tree and generates some rules according to information entropy.
出处
《兰州大学学报(自然科学版)》
CAS
CSCD
北大核心
2007年第6期66-69,共4页
Journal of Lanzhou University(Natural Sciences)
基金
国家自然科学基金(60473045)
河北省科技攻关基金(06213548)资助项目.
关键词
决策树
ID3算法
信息熵
划分
decision tree
ID3 algorithm
information entropy
partition