关键词: 决策森林, 特定信息增益, 数据挖掘

Abstract: Traditional decision tree is based on the information gain of the decision attribute，but sometimes the information gain is changing dynamically according to different values of the decision attribute.This paper considers the differences of the information gain which comes from the different values of the decision attribute and builds the decision forest based on specific information gain.Experiment shows that it has higher classification precision than ID3 though it has a more complex computational procedure.

Key words: decision forest, information gain, data mining