Fig 1.
The trend, seasonality and residual of CGR from 2007 to 2022.
Fig 2.
CGR demonstrates a strong positive correlation (0.90) with SPY during the 2008 Lehman crisis, highlighting its ability to capture market sentiment under volatile conditions.
Fig 3.
The CGR can be viewed an a prediction indicator for the two sudden falls of SPY in the 2008 Financial Crisis.
Fig 4.
Positive correlation of 5-day lagged CGR with SPY during the 2009 European Debt Crisis.
Fig 5.
Positive correlation of 5-day lagged CGR with SPY during the Trade War Fears.
Fig 6.
Positive correlation of 5-day lagged CGR with SPY during the 2018 Fall Selloff.
Fig 7.
The CGR can be viewed an a prediction indicator for the two sudden falls of SPY in the 2018 Fall Selloff.
Fig 8.
SPY price movements during a stable market period (2007–2008).
The steady price trends highlight the absence of significant volatility, establishing this period as a baseline for correlation analysis.
Fig 9.
Correlation matrix of CGR and SPY during a stable market period (2007–2008).
The reduced correlation highlights CGR’s limited applicability in calm market conditions.
Table 1.
The average and maximum values of the VIX during each crisis period are shown alongside the correlation between SPY and CGR.
Table 2.
Hyperparameters for DLSTM network.
Fig 10.
The prediction SPY price on test set.
Table 3.
Hyperparameters for LSTM network.
Table 4.
Directional accuracy of DLSTM, LSTM, and binary classification models.
Fig 11.
Directional predictions on test set.