中国金融风险预警系统的构建研究——基于K-均值聚类算法和BP神经网络  

Research on Early Warning System of Financial Risk in China-Based on K-means Clustering Algorithm and BP Neural Network

作　　者：李梦雨[1] 

机构地区：[1]南开大学经济学院,天津300071

出　　处：《中央财经大学学报》2012年第10期25-30,共6页Journal of Central University of Finance & Economics

基　　金：实现经济金融平稳发展的应对国际金融风险对策研究(08AJY021)

摘　　要：本文首先通过研究我国1994—2011年的经济数据,对关系到金融系统稳定的16项经济变量进行主成分分析,进而将所选变量归结为宏观经济、金融体系、对外经贸三个方面。在此基础上运用K—均值聚类算法,把金融系统风险状态分为四类。继而借助BP神经网络建立了我国金融系统风险的预警模型,并通过2011年的数据对我国2012年金融系统运行状况进行了预测。预测结果表明我国2012年处于轻度风险状态,总需求的回落和资产泡沫的收缩将是影响我国金融系统稳定运行的主要问题。最后对我国如何预测并防范金融风险给出了政策建议。Based on the analysis of economic statistics in China from 1994 to 2011, this article applies prin- cipal components analysis to 16 economic variables related to the stability of finance in order to compartmentalize these variables into macroeconomics, financial system and international business and economics. According to the result of K - means clustering algorithm, we divide financial risk into 4 categories. Then the early warning model of financial system is established by BP Neural Network and the forecast of financial risk in 2012 is conducted u- sing the statistics in 2011. The result demonstrates that the financial risk of China belongs to mild risk status and the main problems which influence the stability of finance is the decline on total demand and the shrink on asset bubbles. Finally policy recommendations are proposed in how to predict and prevent financial risk.

关 键 词：金融风险预警系统 主成分分析K-均值聚类算法BP神经网络 

分 类 号：F830.99[经济管理—金融学]