财经论丛 ›› 2019, Vol. 35 ›› Issue (8): 53-62.

• 金融与投资 • 上一篇    下一篇

互联网金融空间聚集分析及系统性风险防范——基于t-SNE机器学习模型

米传民, 徐润捷, 陶静   

  1. 南京航空航天大学经济与管理学院,江苏 南京 210016
  • 收稿日期:2019-01-23 出版日期:2019-08-10 发布日期:2019-08-28
  • 作者简介:米传民(1976-),男,山东聊城人,南京航空航天大学经济与管理学院副教授,博士;徐润捷(1999-),男,安徽六安人,南京航空航天大学经济与管理学院学生;陶静(1983-),女,安徽寿县人,南京航空航天大学经济与管理学院博士生。
  • 基金资助:
    国家社会科学基金资助项目(17BGL055)

Analysis of Internet Financial Spatial Aggregation and Systematic Risk Prevention——Based on t-SNE Machine Learning Model

MI Chuanmin, XU Runjie, TAO Jing   

  1. College of Economics and Management, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China
  • Received:2019-01-23 Online:2019-08-10 Published:2019-08-28

摘要: 互联网金融同传统金融具有不同的空间聚集特征。互联网金融在带来金融开放、门槛降低、效率提升、成本下降的同时,也给互联网金融体系、乃至整个金融系统带来风险新问题。大量研究表明互联网金融在宏观经济冲击、内部脆弱性等影响下,往往具有与以往不同的系统性金融风险特征。本文利用北京大学的31个省和335个地市区域的互联网金融发展指数有关数据,运用t-SNE机器学习模型进行我国互联网金融发展的降维和聚类分析,得到我国互联网金融空间聚集和不同业务模式发展的分布特征,发现在区域发展程度上存在尖峰厚尾,在业务模式上存在不均衡现象。基于此,提出了考虑互联网金融发展区域差异造成的三方面系统性风险,并为防范互联网金融系统性风险提出建议。

关键词: 互联网金融, 系统性风险, 降维聚类, t-SNE算法

Abstract: The development of Internet finance has different spatial focusing characteristics from traditional finance. Internet finance gives rise to the new problem of systemic risk in the Internet financial system, and even in the financial system as a whole, while bringing about advantages of financial opening, lowering the threshold, improving efficiency and reducing the cost. A large number of studies have shown that Internet finance often has different characteristics of systemic financial risk compared with the traditional finance because of its characteristics such as macroeconomic shock or internal vulnerability. Using the data of Internet financial development in 31 provinces and 335 metropolitan areas, this paper uses t-SNE machine learning algorithm to analyze the health of Internet finance, reveals the distribution characteristics of Internet financial space polymerization and the development of different business models in China, and finds that there are spikes and thick tails in the degree of regional development. There is also an imbalance in the business models. Based on the above analysis, the paper not only suggests that the three aspects of the systemic risk caused by regional differences in Internet financial development should be taken into consideration, but also puts forward some proposals to prevent the systemic risks in Internet finance.

Key words: Internet Finance, Systemic Risk, Dimensionless Clustering, t-SNE Algorithm

中图分类号: