Fuzzy Logic Approach - Computational Intelligence in Time Series Forecasting - page 188

fuzcomb1.m:-fuzzy-combn-of-foreca s t-Vs-Actual-(Tem p -series)

27

26

25

24

23

22

21

20

19

20 40 60 80 100 120 140

time(min)-(nMF=21, training data = 1 to 150

fuzcomb1.m:-fu z zy-combn-of-fore c ast-errors-(Temp- s eries)

0.2

0.15

0.1

0.05

0

-0.05

-0.1

-0.15

-0.2

20 40 60 80 100 120 140

time(min)-( nMF = 21, training data = 1 to 150)

Figure 4.7(a). Nonlinear combination of forecasts using fuzzy model (with training data).
Dots: fuzzy model output; solid line: desired output (upper part), prediction error (bottom)

fuzcomb1.m:-fuz z y-combn-of-foreca s t-Vs-Actual-(Tem p -series)

25

24

23

22

21

20

19

160 170 180 190 200 210 220

time(min)-(nMF=21, training data = 1 to 150

fuzcomb1.m:-f u zzy-combn-of-fore c ast-errors-(Temp- s eries)

0.2

0.15

0.1

0.05

0

-0.05

-0.1

-0.15

-0.2

160 170 180 190 200 210 220

time(min)-( nMF = 21, training data = 1 to 150)

Figure 4.7(b). Nonlinear combination of two forecasts using fuzzy model (with test data).
Dots: fuzzy model output; solid line: desired output (upper part), prediction error (bottom)