改进的ID3算法在个人贷款信用风险评估中的应用

摘要

Based on the model of decision tree classiifcation method of bank loans risk research, the algorithm often tend to attribute values of more, and more value attribute is not necessarily optimal shortcomings, the ID3 was improved, with the average information gain instead of information gain as the basis of selecting the branch attributes, and an evaluation model is established.%文章应用决策树模型分类方法对银行贷款风险进行研究，针对算法往往偏向于取值较多的属性、而取值较多的属性不一定最优的缺点，对ID3进行改进，用平均信息增益代替信息增益作为选取分支属性的依据，进而建立评估模型。