0
$\begingroup$

I was playing around with some data. And I got a distribution which looks like the following. Does this resemble any of the known distributions?

enter image description here

$\endgroup$
  • 1
    $\begingroup$ Lognormal distributions are similar, except they vanish at 0. This might just be a truncated Gaussian. $\endgroup$ – Paul Jun 6 '16 at 11:55
  • $\begingroup$ It is a good idea to log-log plot your histogram so as to get an idea of the decay rate of the tail. If what you get is linear then your tail has a power law; if what you get is concave, then your tail decays faster than polynomially, e.g. exponentially or like a Gaussian. This is hard to see from a bare histogram. (It's also often hard to see from data, because the tail simply doesn't contain that many data points. Still, it's something to try.) $\endgroup$ – Ian Jun 6 '16 at 12:14
  • $\begingroup$ @Paul With the lack of symmetry, it is doubtful that it is a truncated Gaussian, but hopefuly, a Rayleigh distribution... $\endgroup$ – Jean Marie Jun 6 '16 at 13:32
  • $\begingroup$ @JeanMarie rayleigh vanishes at 0, right? $\endgroup$ – dineshdileep Jun 6 '16 at 15:37
  • 1
    $\begingroup$ Yes, a Rayleigh pdf vanishes at 0, but it it is not contradictory with the fact that the initial bin [0, 0.02] collects non-zero values... thus has a non void content. $\endgroup$ – Jean Marie Jun 6 '16 at 15:42
0
$\begingroup$

This could be a Gamma distribution. These distributions contain a factor of the form $x^n e^{-\lambda x}$. Run some through a grapher and see what you think.

$\endgroup$
  • $\begingroup$ The shape of the tail seems somewhat different, no? $\endgroup$ – Ian Jun 6 '16 at 12:11
  • $\begingroup$ @ncmathsadist but won't it vanish at 0? $\endgroup$ – dineshdileep Jun 6 '16 at 15:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.