一种基于离散度的决策树改进算法

在数据挖掘中，决策树方法是一个重点研究方向。ID3方法是著名的决策树算法，在这种算法中，互信息的计算依赖于特征值数目较多的特征，这样不太合理．为此，从离散度的角度，对ID3算法进行改进，通过2种算法的对比实验，证明利用改进后的算法挖掘分类规则，不仅提高了分类的正确率，而且非常高效。

An Improved Decision Tree Algorithm Based on Dispersed Degree

In data mining, decision tree algorithm is a key research direction. ID3 method is a famous decision tree algorithm,among this kind of algorithm, the calculation of information gain depends on the characteristic of characteristic value more in figure,it is not so very rational For this reason, the article improves ID3 algorithm in terms of dispersed degree, through the contrast experiment of two kinds of algorithms, proves that utilizes the algorithm after improving to excavate the categorized rule, not only has improved the correct rate classified, but also very high - efficient.