A Simple Method of Residential Electricity Load Forecasting by Improved Bayesian Neural Networks

Table 12

Performance results of prediction methods.

Method	Corresponding number	MSE	R	MAE	Time (s)
Linear	1	9.90e-02	0.75	0.20	2.10
Interactions Linear	2	9.39e-02	0.77	0.20	4.79
Robust Linear	3	1.05e-01	0.74	0.19	3.74
Stepwise Linear	4	9.40e-02	0.77	0.20	1256.50
Fine Tree	5	1.30e-01	0.68	0.22	7.09
Medium Tree	6	1.09e-01	0.73	0.20	3.53
Coarse Tree	7	1.01e-01	0.75	0.19	2.73
Linear SVM	8	1.03e-01	0.75	0.19	96.97
Quadratic SVM	9	9.78e-02	0.76	0.19	214.74
Cubic SVM	10	9.25e-02	0.77	0.18	828.17
Fine Gaussian SVM	11	2.07e-01	0.49	0.29	159.63
Medium Gaussian SVM	12	9.05e-02	0.78	0.18	112.76
Coarse Gaussian SVM	13	9.88e-02	0.76	0.19	78.12
Boosted Trees	14	9.39e-02	0.77	0.19	23.57
Bagged Trees	15	8.45e-02	0.79	0.17	54.57
BNN_16	16	8.26e-02	0.89	0.18	9.33