基于相空间重构理论的单变量金融时间序列波动预警

摘要

针对单变量金融时间序列，本文基于相空间重构理论和Shannon信息传输理论提出PSR－IDL波动预警方法，对深交所平安银行价格波动序列进行实证。研究发现：在序列接近波动转换临界之前， PSR－IDL指数显著增加并且连续突破警戒阈值； PSR－IDL在接近波动转换临界点之前100天里有74天发出清晰预警信号，但经典的波动预警指标CSD却没有发出清晰稳定的预警信号。这表明，在缺少有效预测模型的情况下， PSR－IDL用作单变量金融时间序列波动临界转换预警是可行有效的。%For the univariate financial time series , the paper proposed an early -warning method based on the PSR -IDL according to the theory of phase space reconstruction and Shannon information transmission and used it to carry out an empirical study on the daily closing stock price of Ping′an bank in the Shenzhen stock exchange market .The results showed the numerical value of PSR -IDL significantly increased and continuously exceeded the warning threshold value before the volatility series reached to the critical transition point .Experimental results showed that the clear warning pre-ceded the critical transition point of volatility series by 100 trade days and accumulatively appeared for 74 days, while the classical early-warning index CSD could not deliver clear and stable signals for early warning .The study provided a proof of principle that the PSR -IDL may be used as an early -warning signal for the volatility critical transition of univa-riate financial time series even in the absence of a predictive model .