Photovoltaic Power Generation Forecasting Using a Novel Hybrid Intelligent Model in Smart Grid

Table 3

Comparative study with literature.

Model	MSE	RMSE	MAPE (%)	MAE	R	Authors
ANN model	90.8078	9.5293	95.65	3.254	0.8511	[28]
Artificial neural networks and an analog ensemble	85.236	9.2323	92.32	3.122	0.8724	[31]
Wavelet transform model	65.2585	8.0782	75.44	2.855	0.9133	[32]
Genetic algorithm optimized hidden Markov model	52.6982	7.6358	65.47	2.544	0.9375	[42]
Multidirectional search optimization algorithm	47.5442	6.8952	61.02	2.142	0.9524	[43]
Multiheaded convolutional neural networks	39.2515	6.2651	54.25	1.857	0.9755	[44]
ANFIS model	34.6934	5.8901	44.23	1.566	0.9814	[45]
ELM model	30.4836	5.5212	31.05	—	0.9824	[47]
ANFIS-PSO model	17.2225	4.1525	26.17	1.356	0.9857	[48]
Nonlinear autoregressive neural network	—	—	—	0.94	—	[49]
The hybrid ACO and PSO model	—	—	3.513	—	—	[50]
ANN-SVM-PSO model	14.9721	3.8693	3.32	0.867	0.9984	Writers