摘要
权益市场的情绪指数对风险预警具有重要的作用,本文提出了一种基于监督学习的沪深300指数多周期情绪指数构建方法,该方法将去除长周期趋势项和短周期趋势的沪深300指数作为长周期风险代理变量和短周期风险代理变量;然后通过严格的因子筛选流程,从因子库中筛选出显著影响长周期风险代理变量和短周期风险代理变量的因子子集,接着利用相关系数加权平均的方法构建长周期情绪指数和短周期情绪指数。实证结果表明,沪深300指数的长周期情绪指数和短周期情绪指数可以有效地表征沪深300指数的长期风险和短期风险程度。利用沪深300多周期情绪指数可有效指导资产配置和风险预警,具有一定的实用价值。
The sentiment index of the equity market plays an important role in risk warning.This paper proposes a method for constructing a multi-period sentiment index of the CSI 300 index based on supervised learning.The method uses the CSI 300 index with long-term trend and shortterm trend removed as the proxies for long-term and short-term risk,respectively.Then,through a rigorous factor screening process,a subset of factors that significantly affect the proxies for long-term and short-term risk is selected from the factor library.The long-term sentiment index and short-term sentiment index are constructed using the method of correlation coefficient weighted average.The empirical results show that the long-term and short-term sentiment indices of the CSI 300 index can effectively characterize the long-term and short-term risk levels of the index.The use of the multiperiod sentiment index of the CSI 300 index can effectively guide asset allocation and risk warning,and has certain practical value.
作者
刘欣元
赵禹平
张京屹
王艺滢
徐建程
LIU Xinyuan;ZHAO Yuping;ZHANG Jingyi;Wang Yiying;XU Jiancheng
出处
《金融发展》
2024年第1期79-88,共10页
Financial Development
