Some related works for cluing to your query are :https://de.mathworks.com/help/fuzzy/anfis.html

http://www.trjfas.org/uploads/pdf_1278.pdf

https://de.mathworks.com/help/fuzzy/anfis.html

Enjoy it and good luck for your modelling.

http://downloads.hindawi.com/journals/ace/2018/8971079.pdf

Article Inference of Water Quality Index Using ANFIA and PCA

Article Performance Evaluation of Two ANFIS Models for Predicting Wa...

https://www.igi-global.com/ViewTitle.aspx?TitleId=233219&isxn=9781522591757

You Matlab Code not answer for your research

My code not worked :(

%ADAPTATIVE NEURO-FUZZY INFERENCE SYSTEM (ANFIS) for Regression

%Description: Implementation of ANFIS using Stochastic Gradient Descent

%for regression problems with instantaneous visualization of the evolution

%of the output, error and membership functions along the epochs.

close all

clear all

clc

%The parameter K defines the number of clusters of fuzzy c-means algorithm

%and consequently the number of fuzzy rules of ANFIS)

%define number of clusters (and rules)

K = 10;

%Define learning rate

alpha = 0.0001;

%Define the maximum number of epochs

maxEpocas = 500;

%dados = load('gasfurnace.txt');

%X1 = dados(:,1);

%Y1 = dados(:,2);

%X1 = [X1;X1];

%Y1 = [Y1;Y1];

X1 = (0:0.05:6*pi)';

Y1 = sin(X1) + 0.3*randn(size(X1));

%X_aug = lagmatrix(X1,[0,1,2,3,4]);

%Y_aug = lagmatrix(Y1,[0,1]);

%X_aug2 = [X_aug(5:end,:),Y_aug(5:end,2)];

%X1 = X_aug2;

%X1 = zscore(X1);

%Y1 = Y1(5:end,:);

y1 = Y1;

tam = round(0.7 * size(X1,1));

idxTreino = [ones(tam,1);zeros(size(X1,1)-tam,1)];

idxTest = [zeros(tam,1);ones(size(X1,1)-tam,1)];

trIdx = logical(idxTreino);

teIdx = logical(idxTest);

X_train = X1(trIdx,:);

Y_train = Y1(trIdx,:);

X_test = X1(teIdx,:);

Y_test = Y1(teIdx,:);

y1 = y1(trIdx);

ErroVal = [];

ErroTreino = [];

X = X_train;

Y = Y_train;

% Fuzzy K-Means Algorithm

%Initialization of the membership matrix.

n = size(X,1);

d = size(X,2);

U = rand(n,K);

idx = zeros(n,1);

U_bin = zeros(n,K);

for i=1:n

U(i,:) = U(i,:)/sum(U(i,:));

[val,cl] = max(U(i,:));

idx(i) = cl;

U_bin(i,idx(i))= 1;

end

initData = [U,idx];

idx_init = idx;

%Initialization of centroids

m = d;

Centroids = zeros(K,d);

oldIdx = idx;

iter = 1;

display('press Ctrl + C to stop')

%Start Looping

while(1)

for i=1:K

indexes = find(idx==i);

Cusj = X(indexes,:);

Centroids(i,:) = ((U(:,i).^m)' * X)./sum(U(:,i).^m);

end

%Calculate distance between points and centroids

for i=1:K

for j=1:n

d(j,i) = 1 / norm(X(j,:) - Centroids(i,:));

end

end

%update membership matrix

for i=1:K

for j=1:n

U(j,i) = d(j,i).^2 / sum(d(j,:).^2);

end

end

%assign each point to the group with it has max membership

for i=1:n

[val,cl] = max(U(i,:));

idx(i) = cl;

end

if isequal(idx,oldIdx),

break;

else

oldIdx = idx;

end;

iter = iter + 1;

end

Xmax = max(X);

Xmin = min(X);

%Parameter initialization

nRegras = K; %Number of rules

m = nRegras;

%n = nObs e d = nDimensions

[n,d] = size(X);