基于ID3算法的道路客运行业信用评级模型

当前信用评级普遍采用的传统方法如专家法、信用评分法中,属性的权重大小受评价者主观影响大,可能造成评价结果偏差失准,并增加评价的时间消耗。本文将决策树模型引入信用评级体系中,在道路客运行业信用评级体系中建立一种新的评价方法。结果表明,采用ID3算法的决策树方法通过定量计算训练集样本中各属性互信息的大小,可以迅速地得到一棵决策树,从而通过对训练样本的统计计算避免了主观判断属性权重带来的弊端。当该决策树经过全面、良好的测试和调整后,可足够健壮,能较好地运用于信用评价。

Credit Rating Model for Road Passenger Transit Industry Based on ID3 Algorithm

There are two disadvantages for traditional methods for credit ratings such as exqert method,credit scoring.The attribute weights may be distorted by the decision-maker,and the process is quite time-consuming.To avoid the drawback,in this paper the author tries to introduce a decision tree method based on ID3 algorithm into Credit Ratings in Road Passenger Transit Industry.The result shows that a decision tree can be quickly built by calculating each attribute's information gain for each sample in a test set,and it can effectively avoid subjective affection.After the decision tree is well tested and adjusted,it can be strong enough and suitable in the use for credit ratings.