摘要
本文通过对信用评价方法中的统计分析方法及机器学习方法进行比较,根据评价指标体系数据的适用性要求,最终选择决策树及其组合方法对P2P信贷个人信用进行评价。通过对模型的准确性及稳定性的综合分析,对本文数据拟合最好的分类方法是随机森林分类方法。
This article compares the statistical analysis methods and the machine learning methods in the credit evaluation field, and according to the applicability of the evaluation index system data, this paper finally uses decision tree and its combination algorithm to evaluate individual credit in P2P lending. Through the comprehensive analysis of the accuracy and stability of the model, the best classification method for the data is random forest classifier.
出处
《统计学与应用》
2017年第3期292-297,共6页
Statistical and Application