A Hybrid Forecasting Model Based on Bivariate Division and a Backpropagation Artificial Neural Network Optimized by Chaos Particle Swarm Optimization for Day-Ahead Electricity Price

Table 9

Models evaluation.

Season	Evaluation items	b-BPANN	CPSO-BPANN	BD-BPANN	CPSO-BD-BPANN
Spring	Accuracy (MAPE)	18.86%	16.81%	12.68%	12.63%
	Calculating time (seconds)	3.25	88.36	6.81	162.29
	Index of stability	6	2	0	0
Summer	Accuracy (MAPE)	11.26%	11.46%	10.28%	9.67%
	Calculating time (seconds)	3.41	88.03	6.92	161.35
	Index of stability	1	0	0	0
Autumn	Accuracy (MAPE)	11.27%	11.37%	10.72%	10.38%
	Calculating time (seconds)	3.45	88.98	6.99	168.89
	Index of stability	1	1	0	0
Winter	Accuracy (MAPE)	8.61%	7.99%	7.74%	7.40%
	Calculating time (seconds)	3.25	84.36	6.49	159.01
	Index of stability	4	2	1	1
Average	Accuracy (MAPE)	12.50%	11.91%	10.36%	10.02%
	Calculating time (seconds)	3.34	87.4325	6.8025	162.885
	Index of stability	3	1.25	0.25	0.25