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

Table 2

Prediction errors of combined EMD models.

Method	Metric	CO2 concentration	Atmospheric pressure	Light intensity	Humidity	Temperature
EMD-RNN	MSE	569.583	792.121	395241.852	1.562	0.703
	RMSE	23.866	28.145	628.683	1.250	0.838
		0.974	0.996	0.958	0.991	0.970
	MAE	13.624	20.050	305.822	0.871	0.644
	MAPE	2.108%	0.021%	——	1.735%	8.158%
	SMAPE	2.147%	0.021%	——	1.721%	7.978%
EMD-CNN	MSE	561.451	753.845	338558.242	1.490	0.533
	RMSE	23.695	27.456	581.858	1.220	0.730
		0.974	0.996	0.959	0.991	0.976
	MAE	13.193	19.215	294.858	0.818	0.542
	MAPE	2.087%	0.020%	——	1.719%	7.853%
	SMAPE	2.072%	0.020%	——	1.687%	7.414%
EMD-LSTM	MSE	553.256	786.858	347852.149	1.375	0.482
	RMSE	23.521	28.051	589.790	1.173	0.694
		0.976	0.996	0.960	0.992	0.978
	MAE	13.188	19.790	298.125	0.771	0.509
	MAPE	2.013%	0.019%	——	1.694%	7.104%
	SMAPE	2.021%	0.019%	——	1.657%	6.916%
EMD-BiLSTM	MSE	532.458	518.857	321129.860	1.145	0.405
	RMSE	23.075	22.778	566.683	1.070	0.636
		0.976	0.997	0.964	0.993	0.986
	MAE	13.103	16.154	277.854	0.708	0.448
	MAPE	1.988%	0.017%	——	1.417%	5.548%
	SMAPE	1.995%	0.017%	——	1.410%	5.124%
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%