Skip to main content
. 2020 Sep 1;212:118743. doi: 10.1016/j.energy.2020.118743

Table 12.

The results of trading hours with the MCS test.

Forecasting models h = 1
h = 5
h = 20
QLIKE
MSE
QLIKE
MSE
QLIKE
MSE
Range SeimQ Range SeimQ Range SeimQ Range SeimQ Range SeimQ Range SeimQ
Panel A: Day (9:00–15:00)
MIDAS-RV 0.6787 0.7369 0.6320 0.6700 0.0826 0.0555 0.0380 0.0230 0.0937 0.1870 0.0322 0.0177
MIDAS-RV-OVX 1.0000 1.0000 1.0000 1.0000 0.0826 0.0532 0.0195 0.0189 0.4651 0.4330 0.0725 0.0697
MS-MIDAS-RV 0.9855 0.9907 0.6322 0.6844 0.2597 0.2597 0.0629 0.0629 0.4651 0.4330 0.8913 0.8913
MS-MIDAS-RV-OVX 0.9855 0.9907 0.9585 0.9585 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000
Panel B: Night (21:00–02:30)
MIDAS-RV 0.5653 0.6203 0.9133 0.9126 1.0000 1.0000 0.9113 0.9113 1.0000 1.0000 0.8987 0.8987
MIDAS-RV-OVX 1.0000 1.0000 0.9133 0.9017 0.0098 0.0675 0.0773 0.2854 0.4021 0.2986 0.3809 0.3236
MS-MIDAS-RV 0.2818 0.4351 1.0000 1.0000 0.2099 0.3994 1.0000 1.0000 0.4021 0.2986 1.0000 1.0000
MS-MIDAS-RV-OVX 0.5653 0.6203 0.8623 0.8455 0.9603 0.9603 0.8107 0.7915 0.4021 0.2907 0.3809 0.1780

Notes: This part shows the results of the models during day-time and over-night trading hours. The forecasts for different horizons h = 1, 5, 20 are shown in different columns, and the results of different trading hours (day and night) are shown in Panel A, B. In this table, the MCS test p-value of 1indicates a better prediction.