Stock Prediction Based on Optimized LSTM and GRU Models

Table 4

Learning results of LASSO-GRU.

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	1293.7802	35.9692	23.8084	831.7062	28.8393	20.3225	117.7369
3	8	20	1444.4125	38.0054	25.5757	910.9886	30.1826	21.6005	249.3114
3	8	30	1123.8232	33.5235	22.7495	826.1986	28.7437	20.0954	402.0947
3	8	40	1482.5836	38.5043	27.8070	943.8668	30.7224	20.9493	508.5391
3	8	50	1280.0657	35.7780	24.5826	978.3127	31.2780	24.1942	657.0113
2	8	10	1377.2985	37.1120	24.9828	834.2753	28.8838	20.0803	90.3936
2	8	20	1136.9731	33.7190	22.8381	799.1834	28.2698	19.5676	200.8435
2	8	30	1245.3809	35.2900	23.5962	864.6091	29.4042	20.8457	276.3479
2	8	40	1500.8726	38.7411	26.3263	1093.2498	33.0643	23.7652	350.2043
2	8	50	1239.8575	35.2116	23.4242	845.8825	29.0841	20.5453	431.4074
3	16	10	1138.9615	33.7485	22.0896	753.7004	27.4536	19.1470	378.2382
3	16	20	1192.8514	34.5377	24.7956	1015.7639	31.8711	23.3095	783.4717
3	16	30	985.8597	31.3984	21.1888	883.4285	29.7225	22.0179	1198.1594
3	16	40	1038.7191	32.2292	21.6453	802.5121	28.3286	19.8618	1622.0595
3	16	50	1116.0420	33.4072	22.7048	792.4473	28.1504	20.0095	2032.0963
2	16	10	1233.5467	35.1219	23.9346	833.0342	28.8623	19.9071	207.6725
2	16	20	1104.4722	33.2336	23.2089	869.4269	29.4860	21.8385	433.1125
2	16	30	1114.1602	33.3790	22.0850	796.8630	28.2288	20.7113	668.5479
2	16	40	1072.4717	32.7486	23.0368	802.4998	28.3284	19.7236	1073.6516
2	16	50	979.3882	31.2952	21.5892	844.8095	29.0656	19.7065	1373.0419
3	32	10	924.5891	30.4071	20.6728	831.0146	28.8273	20.2939	277.7194
3	32	20	862.2056	29.3633	20.4895	1081.1123	32.8803	24.4835	526.1457
3	32	30	907.1031	30.1182	21.2814	1127.4957	33.5782	23.5790	797.8424
3	32	40	847.7834	29.1167	21.0260	1012.2365	31.8157	21.7214	1066.6936
3	32	50	810.5232	28.4697	20.3346	1196.4927	34.5904	24.4721	1349.8547
2	32	10	969.9576	31.1441	20.7674	777.3851	27.8816	19.5336	367.9321
2	32	20	985.3215	31.3898	21.8014	829.0216	28.7927	20.1274	890.5871
2	32	30	861.6357	29.3536	20.5396	952.9534	30.8699	22.5985	1362.8664
2	32	40	1149.7027	33.9073	24.3771	821.3757	28.6597	20.0752	1392.1843
2	32	50	883.3071	29.7205	20.3320	833.8439	28.8764	20.4838	1596.1727

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