多分形互联网金融市场的风险预警模型研究

Research on Risk Early Warning Model of Multi-fractal Internet Financial Market

研究目标:提出基于多分形特征的互联网金融市场风险状态的划分方法,构建非平衡数据集的互联网金融风险预警模型。研究方法:采用日收益率和多分形波动率衡量互联网金融风险并划分风险状态,提出一种利用SMOTEENN采样算法与SVM模型相结合的互联网金融风险预警模型,选取中证互联网金融指数进行实证分析。研究发现:SMOTEENN-SVM模型可以显著地提高SVM模型的预测精度,具有优越的预测性能。研究创新:通过日收益率和多分形波动率刻画互联网金融风险状态,并将SMOTEENN采样算法与SVM模型相结合,建立非平衡样本的互联网金融市场风险预警模型。研究价值:为研究互联网金融风险预警提供新的思路,对防范化解互联网金融风险具有重要的现实意义。

Research Objectives:This paper proposes a method for dividing the risk status of the internet financial market based on multi-fractal characteristics and constructs an internet financial risk early warning model of unbalanced data sets.Research Methods:The rate of return and multi-fractal volatility is used to measure internet financial risks and divide the risk status.This paper constructs an internet financial risk early warning model that combines the SMOTEENN sampling algorithm and the SVM model.The CSI Internet Finance Index is selected for empirical analysis.Research Findings:Compared with single model and SMOTE mixed model,the proposed SMOTEENN-SVM model can significantly improve the prediction accuracy of SVM model and has excellent prediction performance.Research Innovations:The rate of return and multi-fractal volatility is used to characterize the risk status of internet finance,and the SMOTEENN sampling algorithm with the SVM model is used to establish a risk early-warning model of internet finance market with unbalanced samples.Research Value:It provides new ideas for the study of internet financial risk early warning and has important value and significance for preventing and resolving internet financial risks.