一种健壮有效的决策树改进模型

Improved model for robust and effective decision tree

提出了一种健壮有效的决策树改进模型AID3。该决策树模型基于经典的ID3决策树模型,在属性的选取上进行了改进。利用引入属性优先关联度参数计算节点的修正信息增益,并选择具有最高修正增益的属性作为当前节点的测试属性。实验表明,AID3决策树模型在提高分类准确率的同时,有效地增强了模型的健壮性。

The robust and effective decision tree improved model AID3 was introduced. The model was based on ID3, but it was improved in attribute selection. The modified information gain of certain node was calculated by using attribution value association parameter, and attributed with the highest information gain was chosen as the test attribute for the current node. The experiment has proved that AID3 decision tree model improved the accuracy and robustness of classification.