Normal distribution problem using histfit
5 visualizzazioni (ultimi 30 giorni)
Mostra commenti meno recenti
Vlatko Milic
il 17 Mag 2018
Commentato: Vlatko Milic
il 17 Mag 2018
Hi MatLab-friends,
I have a question about histfit. My dataset only contains positive (or 0s) values but when I make my histfit, the normal distribution goes from negative values. How can I make the distribution line only consider positive values?
Attaching figure. How I want the distribution to look is seen with the black line.
Thank you in advance.
P.S.
The code I use now is:
F=d_f_t(:,2) %
n_d=histfit(F)
2 Commenti
dpb
il 17 Mag 2018
The distribution you drew doesn't represent the data well at all; the normal is a very bad choice being symmetric and unbounded; for a continuous distribution you might try lognormal.
Risposta accettata
the cyclist
il 17 Mag 2018
If @dpb had made his comment as an answer, I would have upvoted it.
Your data are clearly not normally distributed. Is there a particular reason that you wanted to fit them with a normal?
You can use the histfit function with a different underlying distribution. dpb's suggestion of a lognormal looks like an excellent choice, from a purely empirical point of view (but you may know something about the underlying process that makes some other distribution a better choice). See the documentation for details, but for example ...
n_d = histfit(F,[ ],'lognormal')
should give you something much better than you have now.
0 Commenti
Più risposte (0)
Vedere anche
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!