Neuro-fuzzy Approach - Computational Intelligence in Time Series Forecasting - page 255

TrnfL M T1:SSE- V s-Epoch (initial S S E = 45.3382, Final S SE= 3.0 8 36e-004 in 999 ep o chs

40

30

20

10

0

100 200 300 400 500 600 700 800 900

No of Epochs

TrnfLMT1 with normalized W ang Data: Neuro-fuzzy output-Vs-Actual , M =10 Rules, 10 GMFs

1.2

1

0.8

0.6

0.4

0.2

0

20 40 60 80 100 120 140 160 180 200

Tim e

Figure 6.9(b). Performance of Takagi-Sugeno-type multi-input single-output neuro-fuzzy
network with M = 10 rules (first model) for normalized Wang data when trained with
proposed Levenberg-Marquardt algorithm.

TS type MISO NF o up ut-vs-actual output (With full-scale Wa n g data), M = 1 0 Rules, 10 G MFs

1

0.5

0

-0.5

-1

-1.5

0 50 100 150 200 250 300 350 400

TS type MISO NF output error (full-scale error) of Wang data

0.02

0.01

0

-0.01

-0.02

0 50 100 150 200 250 300 350 400

Figure 6.9(c). Prediction performance of Takagi-Sugeno-type multi-input single-output
neuro-fuzzy network with M = 10 rules (first model) for non-scaled Wang data after training
with proposed Levenberg-Marquardt algorithm.