基于SVM算法的个人信用评估方法的完善  

Perfection of Personal Credit Evaluation Method Based on SVM Algorithm

作　　者：黄巍[1] 张靓[2] 唐友[1] 

机构地区：[1]黑龙江财经学院,哈尔滨150025 [2]燕京理工学院

出　　处：《黑龙江八一农垦大学学报》2016年第2期105-110,114,共7页journal of heilongjiang bayi agricultural university

摘　　要：在众多的模式识别工具中,支持向量机(Support Vector Machine,SVM)是一种非常有效的解决工具。提出了基于SVM模型提升金融机构对个人信用评估效率的方法。通过对某银行的用户信用数据进行的研究,设计具体评估流程,利用SVM的SMO算法处理参数优化来构建模型,特点是分类精度高、误判率低,具有较好的稳健性,以此来控制消费信贷风险具有良好的适用性。处理商业银行划分信贷等级,应用此种模式可以解决信贷申请和政策实现,具有一定的实际意义。Among many pattern recognition tools,Support Vector Machine（Support Vector Machine,SVM）is a very effective one. A model based on SVM was proposed to promote efficiency of financial institution of personal credit evaluation method. Through researching user credit data of a bank,designing the specific evaluation process,using the SVM SMO algorithm to build the model of processing parameter optimization,it was characterized by high classification precision and low misjudgment rate,so it was stable and could control the consumption credit risk. Dealing with commercial bank credit rating,application of this model would solve the credit application and policy implementation,which had a certain practical significance.

关 键 词：SVM 个人信用评估 SMO算法 

分 类 号：F832.479[经济管理—金融学]