load fuzex1trnData.dat
fis = anfis(fuzex1trnData);
x = fuzex1trnData(:,1);
anfisOutput = evalfis(x,fis);
plot(x,fuzex1trnData(:,2),'*r',x,anfisOutput,'.b')
legend('Training Data','ANFIS Output','Location','NorthWest')
opt = anfisOptions('InitialFIS',4,'EpochNumber',40);
opt.DisplayErrorValues = 0;
opt.DisplayStepSize = 0;
fis = anfis(fuzex1trnData,opt);
figure
anfisOutput = evalfis(x,fis);
plot(x,fuzex1trnData(:,2),'*r',x,anfisOutput,'.b')
legend('Training Data','ANFIS Output','Location','NorthWest')
The above code is just a module to load data and plot it. You need to specify the objective fucntion to be optimized. There has been several works to optimize Adaptive neuro-fuzzy inference system using Genetic Algorithm.
May be the following links can lead you to the solution and help you out:
https://in.mathworks.com/matlabcentral/answers/100323-how-can-i-use-the-genetic-algorithm-ga-to-train-a-neural-network-in-neural-network-toolbox
https://ieeexplore.ieee.org/document/7294321
Article A hybrid of adaptive neuro-fuzzy inference system and geneti...
Dear Malavya,
You can define the mean square error (MSE) between the targets and ANFIS products as the cost function. Alternatively, RMSE, MAE, etc. can be used.
Dear Dr.Amit Rauniyar ,
In case of ANN, it is easy to generate a fitness function for further optimization by using 'genFunction' command. But I could not find any commands in case of ANFIS.
- Badges
- Science topic