基于改进SMOTE的小额贷款公司客户信用风险非均衡SVM分类.docxVIP

基于改进SMOTE的小额贷款公司客户信 用风险非均衡SVM分类 衣柏衡朱建军李杰 南京航空航天大学经济与管理学院 摘要： 研究了小额贷款公司对客户进行信用风险评估时面临的问题，构建了信用风险评 估指标体系，改进了支持向量机(Support Vector Machine, SVM)对非均衡样本分 类时分类超平面偏移的不足。首先分析小额贷款公司业务区域性强、信用数据来 源不规范、评价标准不一致等特点，给出用于客户信用风险评估的四个维度指标。 针对传统SMOTE算法在处理非均衡数据吋对全部少数类样本操作的问题，提出仅 对错分样本人工合成的改进思想,给出具体算法步骤。将改进算法用于某小额贷 款公司客户信用风险评估案例中，分类精确度较其他算法有所提升，表明该方法 的可行性和有效性。 关键词： 小额贷款;信用风险;支持向量机;非均衡数据;SMOTE; 作者简介：衣柏衡(1990-),男(汉族)，天津人，南京航空航天大学经济与管理学院 硕士研究生,研究方向：数据挖掘、系统分析与决策,E-mail :ysb900818@126. com. 收稿日期：2015-05-30 基金：国家社会科学基金重点项目(14AZD049) Imbalanced Data Classification on Micro-Credit Company Customer Credit Risk Assessment Using Improved SMOTE Support Vector Machine YI Bai—heng ZHU Jian-jun LI Jie School of Economics and Management, Nanjing University of Aeronautics and Astronautics; Abstract： A great number of machine learning methods have been successfully applied for customer credit risk assessment cases,and support vector machine(SVM)is considered as an “off-the-shelf” supervised learning algorithm to solve classification problem by many researchers. U nf ortu neitcly，SVM fails to provide excel lent enough classification performance when the data set is imbalanced, i. e., the accuracy of the majority class is usually much higher than that of the minority class due to the shifting of the hyper-plane. In most cases, people pay more attention on the mi no ri ty class such as faul t diag nosis and cred it default. Thus, a Synthctic Minority Ovcr-sampling Technique(SMOTE)is presented to deal with the imbalanced classification by generating new samples in the whole minority class. However, in the process of solving SVM by Sequential Minimal Optimization(SMO)algorithm, only those support vector samples xiwi th the corresponding a i 0can affect the position of the hyper-planc while the samples far from the hypcr-planc have no influence on the final result. It is obvious that the classic SMOTE algorithm can generate more redundant samples which are far from the hyper-plane. In this article, an improved