Fitting distribution

Asked Mar 21 '16 at 00:55

Active Sep 08 '17 at 06:37

Viewed 52 times

I have the following distribution of points and I need to fit a theoretical distribution. It looks like Gamma but I am not entirely sure. Do I have to fit several of them (like Exponential, Poisson etc.) to rule them out or is there a generalized distribution that I can try fitting into that will tell me what is the closest distribution. Something like a generalized normal that can converge on the right type of normal. Generalized normal is clearly ruled out here.

And what would be your best guess for this data. Thanks. Sorry, I couldn't attach an image as don't have required the number of points.

asked Mar 21 '16 at 00:55

user2167741

I would suggest to move this question to stats.stackexchange.com/, however not sure if somebody would want to ask such broad question. – iiivooo Mar 21 '16 at 01:14
Do you need it to be a parameterized distribution? – jonem Mar 21 '16 at 01:29
Yes, preferably. But are there any other options? Thanks! – user2167741 Mar 21 '16 at 01:37
negatively skewed probability distribution – Narasimham Mar 21 '16 at 15:13
Thanks! Can Gamma be a good approximation? – user2167741 Mar 21 '16 at 17:24
Looks more like Extreme Value Distribution. If you fit with maximum likelihood then you can compare the likelihood of different distributions. (But beware distributions with more parameters will fit better so its not just a case of choosing the most likely.) – user121049 Sep 08 '17 at 09:17

Fitting distribution

0 Answers0