0

I need to use fitdistrplus but I have problem with distinguishing when it is discrete and when continuous distribution. Suppose I have

X = {1, 2, 3, 4, 1, 2, 5} # discrete distribution
Y = {1.23, 1.55, 6.45, 3.33, 4.56, 3.33} # continuos distribution

How do we "tell" fitdistrplus what distribution it is, how do we calculate its parameters after?

heisenberg7584
  • 563
  • 2
  • 10
  • 30
  • The `fitdist` function in the `fitdistrplus` package has an argument, `distr=`, that you use to specify which distribution to fit. There are many different continuous (e.g. normal, uniform, lognormal, etc) and discrete (binomial, poisson, negative binomial, etc) distributions. You will find more information in the manual and vignettes that are included in the `fitdistrplus` package. – dcarlson Nov 13 '19 at 19:05
  • Hi, I looked at package site and documentation as well but can't find any array of possible `dist` parameter options... Where can it be found ? – heisenberg7584 Nov 14 '19 at 12:36
  • There may not be a complete list since people can add them all the time. What distribution do you want to fit? Base R has the ones I listed earlier. The examples shown on the manual page for `fitdist` include norm, gamma, weibull, lnorm, nbinom, pois, unif, exp, cauchy, triang, ZIP, llogis, invweibull, gumbel, and pareto. Did you need something else? – dcarlson Nov 14 '19 at 20:39

0 Answers0