基于聚类平衡算法的预测模型在消费信贷违约预测中的应用  

Application of a Prediction Model Based on Clustering Balance Algorithm in Consumer Credit Default Prediction

作　　者：陆慧[1] LU Hui(College of Information Engineering,Anhui Finance&Trade Vocational College,Hefei 230601,China)

机构地区：[1]安徽财贸职业学院信息工程学院,安徽合肥230601

出　　处：《佳木斯大学学报（自然科学版）》2024年第9期149-152,共4页Journal of Jiamusi University：Natural Science Edition

基　　金：2023年度安徽省高校自然科学研究项目(2023AH052069)。

摘　　要：为了提高目前消费信贷违约预测模型的预测准确率,研究将以聚类平衡算法为基础构建一种新型的消费信贷违约预测模型。在聚类平衡算法的性能对比实验中发现,该算法的平均绝对误差为0.00095,显著低于另外两种算法的收敛时的平均绝对误差。随后对信贷违约预测模型进行性能分析,结果显示,研究提出的消费信贷违约预测模型的预测准确率为92.4%,远优于同类型预测模型。以上结果表明,研究提出的消费信贷违约预测模型的预测准确度优于传统预测模型,具有实用价值。In order to improve the prediction accuracy of current consumer credit default predic-tion models,a new type of consumer credit default prediction model will be constructed based on cluste-ring balance algorithm.In the performance comparison experiment of the clustering balance algorithm,it is found that the Mean absolute error of the algorithm is 0.00095,which is significantly lower than the Mean absolute error of the convergence of the other two algorithms.Subsequently,performance a-nalysis was conducted on the credit default prediction model,and the results showed that the prediction accuracy of the consumer credit default prediction model proposed in the study was 92.4%,far superior to similar prediction models.The above results indicate that the prediction accuracy of the consumer credit default prediction model proposed in the study is superior to traditional prediction models,and has practical value.

关 键 词：聚类平衡算法 预测模型 消费信贷 节约成本 

分 类 号：C812[社会学—统计学]