Explanatory Optimization of the Prediction Model for Building Energy Consumption

Table 1

Prediction errors of different models.

Period	Metric	Dataset
		Model 1	Model 2	Model 3	Model 4	Model 5	Our model
January	MAPE	16.24%	14.73%	12.15%	10.58%	13.96%	6.48%
	MSE	7.19	6.33	5.24	4.35	3.85	2.39
February	MAPE	17.54%	15.21%	14.62%	9.58%	9.16%	5.37%
	MSE	7.98	6.08	5.03	4.81	4.59	2.74
March	MAPE	16.61%	14.73%	12.54%	9.38%	14.71%	6.85%
	MSE	7.49	6.63	5.06	4.71	4.38	2.74
April	MAPE	17.28%	16.11%	13.62%	11.48%	9.62%	5.37%
	MSE	6.99	6.52	6.09	4.35	4.18	2.63
May	MAPE	17.84%	17.03%	16.95%	11.41%	13.64%	6.17%
	MSE	7.84	5.87	5.16	4.95	4.35	2.81
Mean	MAPE	15.78%	114.57%	13.05%	11.62%	13.84%	6.37%

MAPE is short for mean absolute percentage error.