2
$\begingroup$

I used the fitdist function from the "fitdistrplus" package in R to estimate the parameters of my data. Once I get the parameters, I must get the confidence interval, right?

From the same package I found the bootdist function but I didn't understand exactly what the interpretation is and how it works and how to get the confidence interval. Especially what does the summary of an object from bootdist give me?

I need someone to tell me how to get the confidence interval of my parameter estimation.

$\endgroup$

1 Answer 1

1
$\begingroup$

bootdist uses the bootstrap method to quantify uncertainty in the parameters estimated from the data.

Especially what does the summary of an object from bootdist give me?

The summary provides the median and 2.5 and 97.5 percentiles of each parameter. If inferior to the whole number of bootstrap iterations, the number of iterations for which the estimation converges is also printed in the summary. via

$\endgroup$
3
  • $\begingroup$ Can i consider the 2.5 and 97.5 percentiles as a confidence interval? $\endgroup$ Commented Jun 25, 2016 at 6:03
  • $\begingroup$ Did you check the link: inside-r.org/packages/cran/fitdistrplus/docs/bootdist? Please do. Especially the part where the output is discussed. Haeds-up: "CI" stands for "confidence interval" $\endgroup$
    – Kostia
    Commented Jun 25, 2016 at 6:07
  • 2
    $\begingroup$ @Kostia "On each bootstrap sample the function mledist (or mmedist, qmedist, mgedist according to the component f$method of the object of class "fitdist") is used to estimate bootstrapped values of parameters. When that function fails to converge, NA values are returned. Medians and 2.5 and 97.5 percentiles are computed by removing NA values." Ouch!! Bad doggie. The fail to converge samples might tend to actually have more extreme values. By tossing these out, the results could be significantly biased. You need a robust solution process to do the bootstrap properly. $\endgroup$ Commented Jun 25, 2016 at 11:46

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.