chi square PDF fitting
4 views (last 30 days)
similar to how I can use fit function with fittype = 'exp1' to find the best f(x)=a*exp(b*x) to match my distribution,
how can I find the best chi square PDF to match my distribution?
By chi square pdf I mean as it defined in: Y = chi2pdf(X,V).
If in exp case I could have the [a,b] that would give me the best fit, here I'm searching for the best V that would give me the best fit.
Jeff Miller on 28 Sep 2018
Cupid will fit the chi-square distribution. For example, if you have a vector of scores, x, from the distribution to be fit, you can get a maximum-likelihood fit with:
mydist = ChiSq(round(mean(x)); % A good starting value.
mydist.EstML(x) % Will adjust V up or down to improve max-likelihood fit.