3
$\begingroup$

This code:

\[ScriptCapitalD] = LogNormalDistribution[-3.3, 0.32];
data = RandomVariate[\[ScriptCapitalD], 1000];
estimated\[ScriptCapitalD] = FindDistribution[data, 5, All]

gives a table of distributions and quality of fit numbers. What is the meaning of the "Internal" and "Score" quality of fit numbers? How are they calculated?

Improve this question
$\endgroup$
1
  • $\begingroup$ The documentation states "The internal information criterion uses a Bayesian information criterion together with priors over TargetFunctions." This doesn't explain much (to me) as it's not clear what priors are used for each distribution. The priors do seem consistent for each distribution considered although the prior seems to be influenced by sample size. In short, I can't explain it. $\endgroup$
    – JimB
    Commented Jan 16, 2018 at 18:55

1 Answer 1

Reset to default
2
$\begingroup$

Mathematica Tech Support said there should only be one column labeled "Internal score." He filed a bug report and also requested documentation to explain the source and meaning of the "Internal score."

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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