汽车消费信贷信用评价模式创新

发布时间：2018-03-15 21:35

本文选题：汽车金融　切入点：汽车消费信贷　出处：《浙江大学》2017年硕士论文　论文类型：学位论文

【摘要】：经历了连续多年世界产销量第一后,我国汽车产业的发展开始步入微增长时代。汽车消费信贷作为推动汽车销量大幅增长的催化剂,越来越受到年轻一代消费群体的接受与认可,但与国外成熟市场相比,仍然具有较大差距。随着90后一代成为新的购车主体,我们认为国内汽车消费信贷必将迎来高速增长,在渗透率上有望快速接近国外成熟市场。而汽车消费信贷市场的进一步成熟,离不开传统信用评价模式的创新与升级。本文通过对现有典型信用评价模型的研究,寻找其中可能存在的缺陷与并深入分析原因。利用Logistic模型与人工神经网络两种方法进行实证对比研究,发现Logistic模型在样本选取、数据处理与录入、以及被拒客户的信息利用等方面的局限性,使得该方法的准确率与稳定性只能在70%左右。因而提出借助较为领先的人工神经网络尝试进行新的信用模型的构建,增加数据采集维度的丰富性、提升处理过程的严密与科学性、注重对大数据的挖掘与发现,从而使新的模型在运算效率、信用评价准确性与稳定性能达到80%以上。我们希望,我们的探索,能够帮助汽车金融机构在汽车消费信贷业务的服务效率与风险控制上均带来一定改善,为我国汽车消费信贷迈向繁荣添砖加瓦。
[Abstract]:After years of production and sales number one in the world, the development of China's automobile industry began to enter the era of micro-growth. It is more and more accepted and recognized by the younger generation of consumers, but there is still a big gap compared with the mature foreign markets. With the post-90s generation becoming the new car buyers, we believe that the domestic auto consumption credit will surely usher in a rapid growth. In terms of penetration rate, it is expected to approach foreign mature market quickly. However, the further maturity of automobile consumption credit market can not be separated from the innovation and upgrading of traditional credit evaluation model. In this paper, the existing typical credit evaluation models are studied. In order to find out the possible defects and analyze the reasons, we use the Logistic model and artificial neural network to carry out the empirical research, and find out that the Logistic model is selected in the sample, data processing and input. Because of the limitation of the information utilization of the rejected customers, the accuracy and stability of the method can only be about 70%. Therefore, a new credit model is proposed with the help of the advanced artificial neural network. To increase the richness of data acquisition dimension, to enhance the strict and scientific processing process, and to pay attention to the mining and discovery of big data, so that the new model can achieve more than 80% operational efficiency, accuracy and stability of credit evaluation. Our exploration can help automobile financial institutions to improve the service efficiency and risk control of auto consumption credit business, and contribute to the prosperity of automobile consumption credit in China.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.479

【参考文献】