Table 4 Accuracy of the WIO-SVR model for the test data.

From: Proposing a hybrid metaheuristic optimization algorithm and machine learning model for energy use forecast in non-residential buildings

Case

Building type

RMSE (kWh)

MAE (kWh)

MAPE (%)

R

Average

Std. dev

Average

Std. dev

Average

Std. dev

Average

Std. dev

1

Commercial building

2.49

0.24

6.25

1.18

6.96

0.72

0.98

0.01

2

Office building

1.35

0.20

1.85

0.53

6.89

1.47

0.95

0.03

3

Authority building

0.31

0.06

0.10

0.04

28.70

13.03

0.78

0.09

4

Office building

1.05

0.15

1.13

0.30

16.86

3.08

0.91

0.10

5

Office building

2.18

0.35

4.85

1.48

10.87

1.74

0.89

0.20

6

Office building

2.06

0.40

4.41

1.57

12.04

5.45

0.85

0.24

7

Authority building

1.00

0.11

1.00

0.19

20.40

2.50

0.90

0.03

8

University building

3.14

0.18

9.91

1.15

32.10

8.64

0.92

0.02

9

University building

3.52

0.25

12.44

1.81

13.26

1.59

0.91

0.02

10

Authority building

0.90

0.29

0.89

0.45

15.18

6.03

0.83

0.14

11

Hospital building

3.63

0.18

13.20

1.33

9.77

0.95

0.92

0.02

12

University building

2.63

0.35

7.01

1.81

12.88

2.91

0.92

0.06