基于偏最小二乘支持向量机模型的个人信用评估研究

Research of personal credit assessment based on partial least squares support vector machine model

为了准确评价客户潜在信用风险,提出了偏最小二乘支持向量机组合评价模型。首先使用偏最小二乘能降低变量间的相关性,支持向量机可用于建立评估模型,然后采用相对误差频率分布作为新的指标评价模型,最后,与常见的评分模型在信用卡数据集上进行了对比。结果表明, PLS-SVM评价模型在有效性、稳定性以及准确性方面均有更好的表现。

In order to accurately identify potential credit risks of customers,a combined evaluation model of partial least squares support vector machine is proposed.First,partial least squares algorithm is used to reduce the correlation between variables,then,support vector machine is combined for modeling.The frequency distribution of relative error is introduced as a new evaluation method.At last,comparative experiments with common method,such as some typical credit scoring models are carried out on the credit card data set.The results show that PLS-SVM model has good performance in accuracy,effectiveness and robustness.