Maximum likelihood fitting for custom function

Question

Lukas Pop 2014-4-9

1
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/125142-maximum-likelihood-fitting-for-custom-function

回答： Chiara Tescione 2021-1-16

I can't find how to fit mle for custom function. mle(data,'pdf',pdf,'cdf',cdf,'start',start,...) doesn't seem appropriate, because it returns probability density values and cumulative density values. I need to fit gev distribution with fixed shape parameter xi to empirical data g; I have following code

xi=-0.1; gev_cdf=@( x,sigma, mu) exp(-(1+xi*((x-mu)/sigma))^(-1/xi));

How can I find maximum likelihood estimates for sigma and mu parameters? Thanks!

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Lukas Pop 2014-4-11

I have found it. The help is tricky, you need to enter pdf, not cdf.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Chiara Tescione 2021-1-16

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/125142-maximum-likelihood-fitting-for-custom-function#answer_599405

Hi, I hope you can help me, I saw that you solved your problem!

How can I estimate the parameters of a function (using an anonymous function) with the maximum likelihood method? I've seen several things on the internet (like mle, fminsearch) but nobody gives me the parameter values!

Example I have:

eq = fittype (@ (a, b, c, x) a * x. ^ 2 + b * x + c);

how do I get the values of a, b and c with the MLE method?

Thanks a lot!