in gaussian membership function from what value of signal the mean is subtracted?

for i=1:n
        A(i)=exp(-0.5*((x-c(i))./s(i)).^2);
end

where

x=sin(2*pi*t)
t=1:.3:10
n=length(x)
c=.4*ones(1,n) %mean
s=.4*ones(1,n) %variance

for each iteration what is the value of x from which mean is subtracted ?

Answer 1

[Image Analyst]

It's all values of x. x is an array and if you don't specify an element, like x(i), then it does ALL elements. So it won't work because you'd be trying to stuff a whole vector into a single element of A. I'd say that you probably want x(i) instead of x.

Try it this vectorized way instead:

t=1:.3:10
x=sin(2*pi*t)
n=length(x)
c=.4*ones(1,n) %mean
s=.4*ones(1,n) %variance
A = exp(-0.5*((x-c)./s).^2)

[pinak parida]

i realize that i have difficulties in viewing an algorithm in matrix or vectorized form, that is why i can not write a correct code in MATLAB.

can you provide some explanatory document with algorithms and programs that enhance my idea about how to proceed and write program in MATLAB.

[Image Analyst]

The t line creates an array of all the t without the need to create the array element by element in a for loop. Then the x lines creates an array of x for every t in the array without using the for loop. Finally A creates an array of that formula for every possible x without using a for loop. I left the semicolons off so you could see what it does.

[pinak parida]

    %fuzzy logic inference system
clc
clear
close all 
% desired signal creation
yd1=-.1;
yd0=-.07;
u3=0;
u2=0;
u1=0;
u0=0;
for t=1:1000
    if t<250
        u(t)=sin(pi*t/25);
    elseif t<500 && t>249
        u(t)=1;
    elseif t<750 && t>499
        u(t)=-1;
    else
        u(t)=0.3*sin(pi*t/25)+0.1*sin(pi*t/32)+0.6*sin(pi*t/10);
    end
end
figure(1)
plot(u)
yd(1)=0.72*yd0+0.025*yd1*u1+0.01*(u2)^2+0.2*u3;
yd(2)=0.72*yd(1)+0.025*yd0*u0+0.01*(u1)^2+0.2*u2;
yd(3)=0.72*yd(2)+0.025*yd(1)*u(1)+0.01*(u0)^2+0.2*u1;
yd(4)=0.72*yd(3)+0.025*yd(2)*u(2)+0.01*(u(1))^2+0.2*u0;
for t=4:999
    yd(t+1)=0.72*yd(t)+0.025*yd(t-1)*u(t-1)+0.01*(u(t-2))^2+0.2*u(t-3);
end
%let us consider 3 input i.e 'x1,x2,x3';so m=j=3
%again each input is fed to 'n=i'number of MF,
%input signal creation
m=3;
n=length(yd);
for j=1:1:m
    for t=1:1000
        if t<250
            x(t)=sin(pi*t/25);
        elseif t<500 && t>249
            x(t)=1;
        elseif t<750 && t>499
            x(t)=-1;
        else
            x(t)=0.3*sin(pi*t/25)+0.1*sin(pi*t/32)+0.6*sin(pi*t/10);
        end
    end
   end 

    figure(1)
plot(x)
U=.1;          %learning parameter
c=0.4*ones(m,n); %mean of guassian mf
s=0.5*ones(m,n); %variance of gaussian mf  
%defining weight
W =0.4*ones(1,n); %weight of LF
% generating MF
q=30;
while(q)
for j=1:1:m
    for i=1:1:n
        %gaussian MF---- A(m,n)=exp(-((x(m)-c(m,n)/s(m,n))^2)
        A(j,:)=exp(-0.5*((x(j)-c(j,:))./s(j,:)).^2);
    end
end

    for i=1:1:n
    for j=1:1:m
        f(i)=min(A(:,i));          %f(i)=min(A(j,i))
        y(i)=sum(W(i)*(x(j)));     %y(i)=sumation j=1 to j=m (w(i)*x(j))
    end
    %output of fuzzy system
    %p=sumation i=1 to i=n (f(i)*y(i))/sumation i=1 to i=n (f(i))
    p(i)=sum(f(i)*y(i))/(sum(f(i))); 
    e(i)=yd(i)-p(i);   %error

    for j=1:1:m
    % updation rule
    W(i) = W(i)+U*e(i)*x(j);
    c(j,:) = c(j,:) + 0.5*e(i)*exp(-(x(j)-c(j,:)./s(j,:)).^2).* (x(j)-c(j,:)./(s(j,:).^2));
    s(j,:) = s(j,:) + 0.5*e(i)*exp(-(x(j)-c(j,:)./s(j,:)).^2).*(x(j)-c(j,:)).^2./((s(j,:)).^3);
end
end
%calculation of mse
p1=(sum(p)/(31));
MSE((31)-q)=sqrt(sum(e.^2)/sum((yd-p1).^2));
q=q-1;
end
figure(3)
plot(yd)
hold on
plot(p,'r')
hold on
plot(e,'g')
figure(4)
plot(MSE,'k','linewidth',2)

sir, i wrote this program for neuro-fuzzy system and got output. the error is decreasing with increase in iteration. BUT with changing the updation equation of mean(c)and variance(s) the output and mse does not change. so i cann't conclude whether program is CORRECT or NOT.

[Image Analyst]

I'm not that familiar with fuzzy algorithms so I wouldn't know if it's correct or not.

[pinak parida]

sir , do you know anyone i can contact?

thankyou

[Image Analyst]

There is a fuzzy logic toolbox, though I don't recall seeing a lot of fuzzy questions here. You can try rephrasing your question, listing "Fuzzy Logic Toolbox" in the product list, and making the title something like "Need help with Fuzzy Logic program" to see if any fuzzy experts here answer you. The Mathworks tech support line won't help with algorithm development, only things like errors, syntax, and other purely MATLAB-related stuff like that.