# How to carry out a chi-squared distribution fit?

14 views (last 30 days)
J.Cam on 25 Mar 2020
Commented: Jeff Miller on 27 Mar 2020
My task is to build a matlab function with 2 inputs: the table of observed data and the level of significance. This function has to compute the chi squared test and deduce which hypothesis to accept.
Now the function is able to compute the chi square but I'm finding difficulty in using the (chi2gof)
function [chi2, df] = Untitled(n)
%% degrees of freedom
r = size(n,1);
c = size(n,2);
df = (r-1)*(c-1)
% Calculate R, C, and T
R = sum(n,2);
C = sum(n);
T = sum(R);
% Calculate \chi^2
CR = (C'*R')';
S = ((n - CR/T).^2)./(CR/T);
chi2 = sum(S(:))
[h,p, stats] = chi2gof(n,'CDF','Alpha',0.05);
end Jeff Miller on 25 Mar 2020
chi2gof is not the function you want here--it is for an entirely different purpose. The last few lines should be something like:
% continued from above...
chi2 = sum(S(:));
p = chi2cdf(chi2,df,'upper');
% Now check p and reject H0 if p < your alpha level (usually 0.05);
% otherwise, do not reject.
% You have already computed everything that would be in stats