Predicting Multidimensional Environmental Factor Trends in Greenhouse Microclimates Using a Hybrid Ensemble Approach

Table 1

Prediction errors of the different models.

Method	Metric	CO2 concentration	Atmospheric pressure	Light intensity	Humidity	Temperature
RNN	MSE	628.711	978.899	487481.897	1.858	0.982
	RMSE	25.074	31.287	698.199	1.363	0.991
		0.973	0.995	0.951	0.991	0.961
	MAE	16.125	22.015	328.486	0.905	0.785
	MAPE	2.253%	0.023%	——	1.811%	8.955%
	SMAPE	2.216%	0.023%	——	1.800%	8.710%
CNN	MSE	592.845	957.548	498524.886	1.784	0.799
	RMSE	24.348	30.944	706.063	1.336	0.894
		0.974	0.996	0.950	0.991	0.970
	MAE	13.953	21.783	330.153	0.888	0.697
	MAPE	2.107%	0.022%	——	1.799%	8.588%
	SMAPE	2.098%	0.022%	——	1.782%	8.220%
LSTM	MSE	607.908	945.915	471187.078	1.667	0.592
	RMSE	24.656	30.756	686.431	1.291	0.769
		0.975	0.996	0.952	0.990	0.975
	MAE	14.242	21.589	325.514	0.846	0.565
	MAPE	2.167%	0.021%	——	1.780%	7.979%
	SMAPE	2.180%	0.021%	——	1.743%	7.444%
BiLSTM	MSE	579.833	696.083	414738.012	1.379	0.468
	RMSE	24.079	26.383	644.002	1.175	0.684
		0.974	0.997	0.953	0.992	0.986
	MAE	13.696	17.867	304.943	0.740	0.474
	MAPE	2.077%	0.018%	——	1.459%	6.259%
	SMAPE	2.056%	0.018%	——	1.460%	5.861%
Informer	MSE	540.253	589.742	339760.369	1.148	0.487
	RMSE	23.243	24.285	582.890	1.071	0.698
		0.975	0.997	0.951	0.993	0.986
	MAE	13.148	16.369	314.202	0.719	0.492
	MAPE	2.061%	0.017%	——	1.397%	6.474%
	SMAPE	2.065%	0.017%	——	1.408%	5.948%
CEEMDAN-Informer	MSE	358.043	242.237	100222.767	0.697	0.249
	RMSE	18.922	15.564	316.580	0.835	0.499
		0.984	0.998	0.986	0.996	0.991
	MAE	11.874	11.098	203.045	0.617	0.346
	MAPE	1.795%	0.011%	——	1.180%	4.250%
	SMAPE	1.795%	0.011%	——	1.185%	4.108%