针对信用评价数据存在离群点和噪声问题,提出一种基于离群点剔除的支持向量机(SVM)信用风险评价模型.该模型利用模糊c-均值聚类算法剔除样本离群点,采用粒子群算法优化支持向量机分类参数,进而提高支持向量机的分类性能.将该方法应用于信用风险评价中的结果表明,相比于其他模型,该方法分类精度更高.
Aiming at the problem of outliers and noise in credit evaluation data, we proposed a support vector machine (SVM) credit risk evaluation model based on eliminating outliers. This model used Fuzzy c-means clustering algorithm to eliminate the outliers. The classification performance of SVM was improved by optimizing the SVM classification parameters by using the particle swarm optimization algorithm. The results of applying the proposed method to the credit risk evaluation show that the classification accuracy is higher than other models.