利用粗糙集和支持向量机的银行借贷风险预测模型

Prediction model of bank lending risk using rough set and support vector machine

将粗糙集和支持向量机相结合,提出利用粗糙集和支持向量机的银行借贷风险预测模型,提高银行个人借贷风险预测精度,为银行判断是否同意借贷提供参考和依据。该模型首先将银行样本数据形成初始决策表,对其进行数据预处理;然后用属性相似度的属性约简算法对决策表进行属性约简,去除冗余属性,获得高效的决策规则;最后将约简后的最小属性作为4种不同核函数的支持向量机学习样本,构建预测模型且进行预测评价对比分析。实验结果表明,使用相同训练样本,高斯核函数的支持向量机预测模型的准确率最高且为93%,说明该预测模型具有较高的精度。通过实验分析,构建预测模型有效地预测了银行借贷风险,且为银行借贷管理提供辅助决策。

This paper proposes a bank lending risk prediction model by combining rough set and support vector machine,in order to improve the accuracy of personal lending risk prediction,and provide reference and basis for banks to judge whether to agree to lend or not.The model firstly forms the initial decision table from the bank sample data and preprocesses the data.Then the attribute reduction algorithm of attribute similarity is used to reduce the attributes of the decision table so as to remove redundant attributes and obtain efficient decision rules.Finally,the minimum attribute after reduction is taken as the learning sample of support vector machine of four different kernel functions to construct the prediction model and carry out comparative analysis of prediction evaluation.Experimental results show that with the same training samples,the accuracy of Gaussian kernel SVM prediction model has a high accuracy(93%).Through the experimental analysis,the prediction model is constructed to predict the risk of bank lending effectively,and provide auxiliary decision for bank lending management.