基于最近邻法和支持向量机的个人信用评估方法

本文针对两类个人信用数据混叠较严重的数据集，提出对数据集先利用最近邻算法进行修剪，再应用SVM算法对个人信用进行评估的NN—SVM方法。仿真实验表明基于NN—SVM算法的个人信用评估方法比直接用SVM算法进行分析来的更加准确。同时，对比RBF_LS—SVM，Linear LS-SVM,Region single tree等算法的结果，发现NN—SVM算法明显优于其它算法。

Based on Nearest Neighbor and Support Vector Machine Method and Its Application in Personal Credit Assessment

In this dissertation, a new method based on Nearest Neighbor and Support Vector Machine （NN-SVM） is presented to deal with credit assessment. Since the hank data that we find blends together, we use NN to shear bank data. And then we use SVM to deal with credit assessment for distinguishing good credits and bad credits. The outcome of test indicates that the NN-SVM is better than SVM. In order to make sure NN-SVM is good enough for this data, we compare NN-SVM with others such like BRF-SVM and Linear LS-SVM, Region single tree, the result shows that NN- SVM is better than others.