Novel Cooperative Multi-Input Multilayer Perceptron Neural Network Performance Analysis with Application of Solar Irradiance Forecasting

Table 6

Five input multilayer perceptron neural network performance analyses with different numbers of hidden neurons.

Number of hidden neurons	MSE	RMSE	MAE	MRE	MAPE		Time
1	8.8866-04	2.9800-02	2.1400-02	9.1578-05	9.2000-03	1	63
2	2.1000-03	4.5800-02	2.8200-02	1.2054-04	1.2100-02	1	77
3	2.1164+01	4.6005	2.5291	1.0800-02	1.0798	9.9997-01	85
4	1.4053-04	1.1900-02	6.8000-03	2.8928-05	2.9000-03	1	86
5	1.8456-04	1.3600-02	7.3000-03	3.1340-05	3.1000-03	1	76
6	1.4665	1.2110	1.0607	4.5000-03	4.5290-01	1	114
7	2.4758-04	1.5700-02	1.1800-02	5.0208-05	5.0000-03	1	78
8	4.3498+02	2.0856+01	1.3909+01	5.9400-02	5.9386	9.9945-01	136
9	7.9277+01	8.9037	6.7448	2.8800-02	2.8798	9.9989-01	144
10	1.1000-03	3.2500-02	2.3300-02	9.9492-05	9.9492-05	1	146
11	1.5882+02	1.2602+01	1.1004+01	4.7000-02	4.6982	9.9984-01	167
12	2.2279+03	4.7200+01	3.5290+01	1.5070-01	1.5068+01	9.9655-01	176
13	5.3100-02	2.3050-01	8.9600-02	3.8240-04	3.8200-02	1	183
14	1.4020-01	3.7450-01	2.2400-01	9.5640-04	9.5600-02	1	192
15	9.0506+01	9.5135	3.7441	1.6000-02	1.5986	9.9960-01	200
16	2.1349+01	4.6205	2.0997	9.0000-03	8.9650-01	9.9994-01	162
17	1.3000-03	3.6500-02	2.8500-02	1.2188-04	1.2200-02	1	124
18	3.0136+02	1.7360+01	5.8302	2.4900-02	2.4893	9.9907-01	100
19	2.1978+03	4.6881+01	3.8933+01	1.6620-01	1.6623+01	9.9662-01	247
20	1.1940+02	1.0927+01	7.6431	3.2600-02	3.2633	9.9984-01	240

Bold implies the optimal result.