Table 4 Sensitivity analysis 2 for identifying the appropriate inputs.

From: Developing a hybrid time-series artificial intelligence model to forecast energy use in buildings

Scenario

Combinations of different inputs for predicting energy consumption

Accuracy indices with the test data

ANC

Rank

RMSE

(kWh)

MAE

(kWh)

MAPE

(%)

R

  

1

Y

1.24

1.53

4.69

0.991

0.88

6

2

Y, HoD, and DoW

1.18

1.40

4.38

0.993

0.70

5

3

Y and T

1.15

1.32

4.18

0.993

0.47

3

4

Y, T, and H

1.16

1.34

4.46

0.993

0.53

4

5

Y, T, HoD, and DoW

1.07

1.14

3.42

0.995

0.00

1

6

Y, T, H, HoD, and DoW

1.08

1.16

3.53

0.995

0.06

2

  1. Note: ANC stands for average normalized accuracy; A scenario with bold numbers indicate its outstanding performance compared to others.
Back to article page