欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
在国际上，个人信用评分是个人信贷风险防范的重要环节，而中国目前还没有制定出一套规范的个人信用评分指标体系和方法。本研究利用真实的个人消费信贷数据，选择适合的字段作为指标变量并进行归一化处理，结合个人信用评分的特点，选择BP神经网络算法建立了个人信用评分模型。实证研究表明：该模型预测精确度较高，具有较强的判别预测能力，但稳健性却不是很理想；适用于样本分布不断变化或数据结构不太清楚的情况，但却存在过度拟合的问题。
英文摘要：
Personal credit scoring is one of the most important indexes of evaluating the personal credit risk used in the world; however, no standard personal credit scoring index system or method has been worked out in China up to now. This paper establishes a personal credit scoring model by using the BP neural-network algorithm and considering the characteristics of personal credit scoring on the basis of some real personal consumption credit data and proper index variables. This case study has proved that the model is very accurate and its forecasting ability is very strong, but its stability is not good, and it is suitable for the cases with a changing distribution of samples or with an unclear data structure, but it has an over-fitting problem.
同期刊论文项目
同项目期刊论文
期刊信息
位置：