Electricity Demand Projection Using a Path-Coefficient Analysis and BAG-SA Approach: A Case Study of China

Table 4

Comparison of results from various electricity demand estimation models.

Models	Optimization algorithms		MAPE of training set	MAPE of testing set
Multiple linear	GA	0.9871	7.0871%	2.9348%
	PSO	0.9908	6.3245%	2.5802%
	BA	0.9929	5.9254%	2.3702%
	BAG-SA	0.9943	5.1729%	1.9288%
Quadratic	GA	0.9883	6.6950%	2.7215%
	PSO	0.9915	6.1841%	2.5241%
	BA	0.9940	5.6614%	2.2646%
	BAG-SA	0.9985	2.7883%	1.1772%