Brent Crude Oil Price Forecast Utilizing Deep Neural Network Architectures

Table 8

Comparison of modeling results.

Parameter	One-layer LSTM with SGDM solver	One-layer LSTM with RMSProp solver	One-layer LSTM with Adam solver	A Bi-LSTM with SGDM solver	A Bi-LSTM with RMSProp solver	Two-layer LSTM with SGDM solver	Three-layer LSTM with SGDM solver
RMSE	1.88	1.59	2.73	2.18	1.94	1.53	1.58
Maximum error	5.825	12.44	17.31	7.74	11.24	7	5
Modeling time	429	63.75	248	822	114	345.32	911
MSE	3.9	3.36	5.76	3.60	3.72	2.68	2.70
Feedback regressors	[1, 2, 3, 4]	[1, 2, 3, 4]	[1, 2, 7]	[1, 2, 3, 4]	[1, 2]	[1, 2]	[1, 2, 3]
Hidden layers	100	10	80	100	10	80	80