论文部分内容阅读
电网建设工程通过项目融资租赁进行快速融资的同时,给租赁公司带来巨大的信用风险.通过事前对承租人进行信用评价,能够有效降低信用风险损失.针对电网企业信用评价的多属性非线性特征,提出了基于独立分量分析技术-支持向量机的信用评价混合模型.首先,采用独立分量分析技术对信用属性数据进行属性重构,实现属性数据的去噪.然后,将重构后的新信用属性数据用于支持向量机的训练建模.最后,通过实例模拟对比分析了独立分量分析技术对支持向量机分类的有效性.结果表明,独立分量分析技术能够改善信用属性数据特征,并且在多属性分类问题中,独立分量分析技术有助于提高支持向量机分类的准确率.
Grid construction project brings a huge credit risk to the leasing company through the project financing leasing, which can effectively reduce the credit risk loss through credit evaluation of the lessee in advance.For the multi-attribute non-linear characteristics of the credit rating of the grid enterprise , A mixed credit evaluation model based on independent component analysis and support vector machine is proposed.Firstly, independent component analysis is used to reconstruct the credit attribute data to realize the denoising of attribute data.And then, the reconstructed credit The attribute data is used to train SVM training model.Finally, the effectiveness of independent component analysis (SVM) for SVM classification is compared and analyzed by example simulation.The results show that ICA can improve the credit attribute data characteristics, In attribute classification, independent component analysis helps to improve the accuracy of SVM classification.