Stock Prediction Based on Optimized LSTM and GRU Models

Table 5

Learning results of PCA-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	1369.6727	37.0091	27.3800	29235.9590	170.9853	152.2538	72.1253
3	8	20	2434.9895	49.3456	35.6683	12723.6592	112.7992	93.6385	168.9686
3	8	30	2101.9541	45.8471	33.8847	34108.7148	184.6854	166.6717	254.5802
3	8	40	2574.5256	50.7398	37.5205	21786.1641	147.6014	127.3568	333.7451
3	8	50	1194.8505	34.5666	25.1312	29938.3730	173.0271	151.4763	458.2760
2	8	10	1196.4231	34.5893	23.0424	10500.0391	102.4697	82.4290	54.2256
2	8	20	2299.9802	47.9581	36.0542	29103.7676	170.5983	151.9113	112.4777
2	8	30	1426.4019	37.7677	26.6629	21631.0195	147.0749	121.6393	167.5188
2	8	40	1835.9618	42.8481	29.4047	9878.1016	99.3886	80.6860	225.1572
2	8	50	1503.7362	38.7780	26.7672	10440.7920	102.1802	86.5779	275.0539
3	16	10	943.2319	30.7121	21.8401	29209.2168	170.9070	148.9597	116.7384
3	16	20	816.0226	28.5661	20.8915	12768.4766	112.9977	99.0165	255.2906
3	16	30	965.2081	31.0678	23.1747	14524.5684	120.5179	99.9182	815.1597
3	16	40	980.0072	31.3051	22.1477	12230.3311	110.5908	84.3423	1096.2136
3	16	50	803.9420	28.3539	20.7994	9752.1914	98.7532	81.7505	1354.4110
2	16	10	776.4768	27.8653	19.3489	7601.5073	87.1866	74.7424	70.5103
2	16	20	784.2062	28.0037	21.0686	18093.4902	134.5120	119.4448	176.9984
2	16	30	747.7553	27.3451	20.0675	8023.9351	89.5764	77.0783	236.3435
2	16	40	720.6558	26.8450	19.3404	18241.6797	135.0617	117.5537	378.0467
2	16	50	713.9094	26.7191	19.0031	9412.2773	97.0169	79.6811	484.1664
3	32	10	453.6670	21.2995	15.6519	6122.3311	78.2453	61.3247	469.7184
3	32	20	456.4868	21.3656	15.9169	11951.0684	109.3209	90.6603	977.3712
3	32	30	626.6472	25.0329	19.3384	11865.4746	108.9288	86.8948	1483.5679
3	32	40	529.3604	23.0078	17.0553	9151.2109	95.6620	76.8583	1991.7930
3	32	50	504.7110	22.4658	16.9049	30787.4473	175.4635	153.6052	3289.7475
2	32	10	425.4210	20.6257	15.0201	6595.3359	81.2117	63.7287	268.0692
2	32	20	653.8412	25.5703	18.7442	10183.2344	100.9120	84.6054	539.5263
2	32	30	508.9430	22.5598	16.2366	6050.9692	77.7880	61.9705	856.6547
2	32	40	630.0575	25.1009	18.9956	7276.7456	85.3038	69.0245	1345.9522
2	32	50	623.7015	24.9740	18.0729	9594.8418	97.9533	80.5411	1678.2588

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