基于代价敏感SVM的个人信用评估模型

The Credit Scoring Model for Individuals Based on Cost-Sensitive SVM Model

摘要: 本文主要研究了客户借贷规模在个人信用评估中的影响，首先按照客户借贷规模对所有借贷客户进行分类，对不同类别的客户的错判所导致的损失赋予不同的权重；然后使用代价敏感支持向量机，构建了一种新的个人信贷评估模型。实证检验说明，此模型可以降低信贷公司在个人信用评估过程中的总损失，同时提高对借贷规模相对较大的客户的判别准确率。

Abstract: In this paper, we mainly focus on the influence of the amount of credit in personal credit scoring. At first, we divided the customers into two parts based on their credit amount, and give different weights to the wrong-judging of different parts of customers; Then we used the cost-sensitive SVM model to establish a new personal credit evaluation model. Results suggest that this new model can decrease the total cost of personal credit evaluation, and improve the accuracy of judgment of those large customers who will break the contract.