0
$\begingroup$

I'm attempting to approximate the variance within how quickly a given housing unit could be re-rented once the previous tenants vacate. Assuming notice is given, it is much more likely that it will be rented again in 0-2 months but there is a slim chance it will extend far longer. Does anyone know how I could discern whether a beta or gamma distribution (or some other option, maybe a standard distribution with skew?) could be better used to model this variance? How would I assess what degrees of freedom to use for the model in Excel?

Thanks!

$\endgroup$
1
  • $\begingroup$ Welcome to math.SE ! This question may be more suited to the statistics stackexchange community. Have you tried to ask it on Cross Validated ? $\endgroup$
    – Tom-Tom
    Feb 5, 2014 at 21:21

1 Answer 1

0
$\begingroup$

I would suggest the gamma family of distributions, which will include the exponential distribution as a special case. Both are right skewed and do not have "degrees of freedom" parameters.

The beta distribution has a finite upper bound, which you may not want if you don't really know what that could be. Also, it has more parameters (3 or 4 depending on if you fix its left bound at 0) to estimate from the data than the exponential (1 parameter) or gamma (2 parameters), so you will waste data estimating extra parameters.

I'd start with the exponential and see if it makes sense with your data, then go to the gamma if you need more flexiblity.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .