Skip to main content

Table 11 Comparison results of NoVaS-type, GARCH(1,1) and GJR-GARCH(1,1) methods on forecasting 500 size simulated Model 8 data

From: A model-free approach to do long-term volatility forecasting and its variants

500 size

GE-NoVaS

GA-NoVaS

P-GA-NoVaS

GARCH-direct

GJR-GARCH

M8-1step

1.08923

1.10182

1.11782

1.36494

1.00000

M8-5steps

1.34595

1.31927

1.37408

3.85497

1.00000

M8-30steps

1.19125

0.97943

1.06189

1125.00581

1.00000

  1. The benchmark is the GJR-GARCH(1,1) method, so numerical values in the table corresponding to this method are 1. Other numerical values are relative performance compared to the GJR-GARCH(1,1) method. This table is also based on the average result of 5 replications