Stock Prediction Based on Optimized LSTM and GRU Models

Table 2

Learning results of LASSO-LSTM.

Number of layers	Number of neurons	Look-back value	Train MSE	Train RMSE	Train MAE	Test MSE	Test RMSE	Test MAE	Train time
3	8	10	1904.5410	43.6410	29.6357	979.393	31.2953	23.7622	132.8780
3	8	20	1272.1884	35.6678	24.2016	904.9543	30.0825	22.6913	307.9990
3	8	30	2142.9192	46.2917	36.9945	957.2454	30.9394	22.6934	430.9928
3	8	40	4629.96388	68.0438	61.9541	5342.1396	73.0899	67.8341	576.7966
3	8	50	1214.7745	34.8536	23.3589	881.0265	29.6821	21.0112	766.2157
2	8	10	1248.6508	35.3363	23.4574	777.5208	27.8841	19.6554	95.1825
2	8	20	1608.0676	40.1007	27.8215	991.3244	31.4853	21.8147	213.3500
2	8	30	1251.1989	35.3723	22.9057	858.0669	29.2928	19.7988	320.6939
2	8	40	1202.1561	34.6721	23.1705	883.7101	29.7273	21.3382	411.6595
2	8	50	1206.7042	34.7376	23.1075	850.2483	29.1590	20.6415	490.2418
3	16	10	1031.0358	32.1097	21.8451	778.9459	27.9096	19.9938	376.8092
3	16	20	6555.5376	80.9663	75.5480	9468.3418	97.3054	93.4085	779.2415
3	16	30	1254.3915	35.4174	23.0273	733.8773	27.0902	18.6517	1189.4983
3	16	40	1015.5285	31.8674	21.3504	872.5764	29.5394	20.9425	1603.3826
3	16	50	984.6891	31.3798	22.6411	892.3263	29.8718	20.5884	2065.8919
2	16	10	1051.3169	32.4240	21.6973	786.5796	28.0460	19.8271	212.8883
2	16	20	1100.5430	33.1744	21.7411	734.0122	27.0927	18.7960	442.2047
2	16	30	1854.9745	43.0694	34.8938	1331.5276	36.4901	30.3015	725.7830
2	16	40	984.7316	31.3804	20.8187	803.8667	28.3525	20.9327	1080.9957
2	16	50	1006.7289	31.7290	21.1893	846.4849	29.0944	21.3891	1346.3720
3	32	10	787.5522	28.0634	20.2811	852.4590	29.1969	20.1759	306.4265
3	32	20	816.3383	28.5716	20.2925	1199.1475	34.6287	23.9940	636.0178
3	32	30	908.5170	30.1416	20.8915	1850.9948	43.0232	30.6120	970.8293
3	32	40	844.3165	29.0571	20.1408	2720.0981	52.1546	32.7257	1291.5366
3	32	50	1057.6416	32.5214	21.6094	2061.0061	45.3983	29.4876	1356.4536
2	32	10	1117.4025	33.4276	22.2174	799.0801	28.2680	19.8247	454.6504
2	32	20	919.8626	30.3292	21.2240	751.9146	27.4211	19.1840	929.6708
2	32	30	1096.1162	33.1076	23.0584	2220.7078	47.1244	31.1358	1358.4529
2	32	40	1124.6395	33.5356	23.5911	1125.7322	33.5519	25.2733	1385.7139
2	32	50	1442.5688	37.9812	29.5802	1449.2459	38.0690	29.8898	1497.0255

The number of epochs is 1000, learning rate is 0.001, and the activation function is .