Logistic回归与分类树模型的比较

为了有效地评估客户的可信度,提高信贷机构经济效益。文中通过统计学中的参数方法Logistic回归和非参数方法分类树这两种方法,建立两种模型对数据进行预测,应用SPSS软件的Binary Logistic Regression方法,利用ROC曲线的性质来对模型的性能进行评价,根据输出结果比较两种模型在应用中都是可行有效的,在实际操作中应因地制宜,把握两种方法的优势,得到更有价值的结果.

Comparison between Logistic Regression and Classification Tree Mode

The study aims to effectively evaluate the customer credibility and improve the economic benefits of credit mechanism .By their methods of Logistic regression and classification tree ,two kinds of models were established to forecast data .Their performance was evaluated by the Binary Logistic Regression method of SPSS software and the properties of the ROC curve .The output results show that both models are feasible and effective .In practice ,the advantages of the two methods should be taken to get more valuable results .