摘要
本文根据ID3算法中信息增益计算原理的特点,利用数学上等价无穷小的性质提出一种新的改进的ID3算法,减少了信息增益的计算量,进而提高ID3算法中信息增益的计算效率。与原ID3算法相比,改进的ID3算法在构造决策树时具有相同的准确率和更高的计算速度。
According to the gain information in the ID3 algorithm calculation principle, we use the mathematical equivalence of the infinitesimal nature of a newly improved ID3 algorithm to reduce the amount of computation gain information, and improve the information gain computational efficiency in the ID3 algorithm. Compared with the original ID3 algorithm, the Improved ID3 algorithm has the same rate of accuracy and higher speed in the decision tree structure.
出处
《计算机工程与科学》
CSCD
北大核心
2009年第6期109-111,共3页
Computer Engineering & Science
基金
2007年娄底职业技术学院科研资助项目(07ZF021)
关键词
决策树
ID3算法
信息增益
等价无穷小
decision tree
ID3 algorithm
information gain
equivalent infinitesimal
