基于资金流的金融系统异常的统计监测

Statistical Monitoring of Financial System Abnormalities Based on SVDD and PCA

金融系统运行的稳健性对国家经济有重要影响。监测金融系统运行中的资金流异常,能够减少金融风险累积,防范金融危机。论文从资金流角度出发,统筹考虑金融系统功能和金融系统功能有效发挥对宏观经济目标的影响,通过结构方程模型构建监测指标体系。结合数据缺少负样本、监测指标的相关性、样本数据的多寡等特点,选取SVDD、PCA等异常检测方法进行金融系统异常监测模型的实证分析,并根据检测结果分析异常出现的原因。

By detecting the abnormal flow of funds in the operation of the financial system, it is possible to reduce the accumulation of financial risks and prevent financial crises. From the perspective of capital flow, this article takes into consideration the overall impact of financial system functions and the effective exertion of financial system functions on macroeconomic objectives and simulates the detection index system through a structural equation model. Finally, based on the characteristics of the data, an empirical analysis of the abnormal monitoring model of the financial system is performed using SVDD, principal component and other anomaly detection methods. They further analyze the cause of the anomaly based on these detection results.