Can you help me fix this?

Im using the following function

function beta = arima(y, p, q, C, sigma)

if nargin < 4
C = 0;
end
if nargin < 5
sigma = 1;
end

y = y(:);
N = length(y);

e = sigma * randn(N, 1);

Y = y - e;

% By = y(1:end-1) y(1:end-2) ... y(:, end-p)
By = arrayfun(@(j) [zeros(j,1); y(1:end-j)], 1:p, 'UniformOutput' , false); By = [By{:}];
Be = arrayfun(@(j) [zeros(j,1); e(1:end-j)], 1:q, 'UniformOutput' , false); Be = [Be{:}];

if C == 0
cvec = [];
else
cvec = ones(N,1);
end

X = [cvec By Be];

x1 = X(any(X,2),:);

beta = lsqnonneg (x1, Y);
end

*In the following code*

a0 = 0.05; a1 = 0.1; b1 = 0.85;

epsi=zeros(3000,1);
simsig=zeros(3000,1);

for i = 1:3000
                 if (i==1) 
                     simsig(i) = a0 ;
                     s=(simsig(i))^0.5;
                     epsi(i) = normrnd(0,1) * s;
                 else
                     simsig(i) = a0+ a1*(epsi(i-1))^2+ b1*simsig(i-1);
                     s=(simsig(i))^0.5;
                     epsi(i) = normrnd(0,1) * s;
                 end
end

yt1=epsi.^2;

beta = arima(yt1, 1, 1, 1, 1);

I expect to get

Constant estimate close to 0.05

ar parameter estimate close to 0.95

and ma parameter estimate close to 0.85

However, I am not getting the results .. Can someone fix this please?

[Walter Roberson]

Which compiler are you using, on which operating system?

Have you tried debugging it at the MATLAB level before generating C / C++ from it?

[dav]

I am doing this in Matlab.. I dont use c/c++ Thanks

[Walter Roberson]

Your tags include "matlab coder" and "code generation", which have to do with generating C or C++ code. Your other tag, "matlab compiler" needs a C or C++ compiler temporarily in order to generate the final executable even though it does not generate C or C++ code for the MATLAB code itself.

[dav]

what tags should I put when I need help with just matlab coding then?

Thanks

[Walter Roberson]

http://www.mathworks.co.uk/matlabcentral/answers/43073-a-guide-to-tags

Answer 1

[Jan]

Of course we cannot fix your function, because we do not know, what you are wanting to calculate. All we have is the code, which might contain a bug anywhere, and the the expected result. But this is not enough information to locate a bug.

Perhaps this simplified questions reveals the core of the problem of your question:

 I have two values a=9 and b=10. I add them: |c = a + b| and get 19. But I want to get 20. Please fix the code.

There is an infinite number of possible modifications and based on the given information, it is not possible to decide for a specific one. Ok?