论文部分内容阅读
针对企业债务违约损失率判别问题中属性变量居多这一特点,选择支持向量机模型进行判别,并从贷款回收的角度将以往简单的两类有无回收模式(有回收和无回收)扩展为三类。为提高模型效率,将逐步判别分析法应用到模型变量的选择上,同时为了避免人为选择参数的随意性,采用粒子群算法优化支持向量机的参数,将建立的PSO-SVM多分类判别模型对500笔银行贷款进行实证研究。结果表明,该模型不仅提高了分类准确率,而且具有良好的稳健性。
In view of the feature of the majority of attribute variables in the judgment of default loss ratio of enterprises, SVM model is selected to discriminate, and from the point of view of loan recovery, the past simple two types of recovery modes (with and without recycling) are expanded to three class. In order to improve the efficiency of the model, the stepwise discriminant analysis method is applied to the selection of model variables. In order to avoid arbitrary choice of parameters, PSO is used to optimize the SVM parameters. The PSO-SVM multi-classification discriminant model 500 bank loans for empirical research. The results show that this model not only improves the classification accuracy, but also has good robustness.