【摘要】 为提高分类算法在信贷风险领域不平衡数据的预测性能,提出一种基于高斯混合模型(Gaussian mixture model,GMM)的欠采样算法,将其应用在信贷不平衡数据领域中。采用高斯混合模型对多数类样本进行聚类欠采样(under-sampling),消除样本间的不平衡问题。实验比较该算法与传统的欠采样方法,进行该算法的抗噪鲁棒性分析,实验结果表明,该算法能够有效提升分类器的性能,其对信贷数据集具有较强的鲁棒性。更多还原

【Abstract】 To improve the prediction performance of classification algorithm for imbalanced data in the field of credit risk,an under-sampling algorithm based on Gauss mixture model(GMM)was proposed and applied in the field of imbalanced data of credit risk.The Gauss mixture model was used to cluster under-sampling for most samples,thus eliminating the imbalance between samples.The traditional under-sampling methods were compared in the experiment,and the robustness to the noise of the algorithm was analyzed.The results show that the proposed method can effectively improve the performance of the classifier,and it has strong robustness to credit data sets.更多还原