In teachingApps: Apps for Teaching Statistics, R Programming, and Shiny App Development

Properties of the Gamma Distribution

• The gamma distribution is a member of the exponential family of the form

$$ \exp\bigg[\frac{-x}{\theta} + (\kappa-1)\log(x) - \log\Big(\Gamma[\kappa]\Big)-\log\Big(\theta^{\kappa}\Big)\bigg] $$

• where

  • $\kappa - 1$ and $-\theta^{-1}$ are the natural parameters

  • $log[x]$ and $x$ are the sufficient statistics

Relationships to Other Distributions

• If $X \sim GAM(\kappa = 1,\theta = \lambda)$, then $X\sim EXP(\lambda)$ where $\lambda$ is a scale parameter

• If $X \sim GAM(\kappa = \nu/2,\theta = 2)$, then $X\sim \chi^{2}(\nu)$ where $\nu$ is the number of degrees of freedom

• If $X \sim GAM(\kappa \in \mathbb{Z},\theta \in \mathbb{R}^{+})$, then $X\sim ERLANG(\kappa,\theta)$ and represents the time of the $\kappa^{th}$ arrival in a Poisson process with intensity $1/\theta$

• If $X \sim GAM(\kappa,\theta)$, then $1/X\sim \mbox{INV-GAMMA}(\kappa,1/\theta)$

• If $X \sim GAM(\kappa,\theta)$, then $X^{\beta}\sim GENG(\kappa,\beta,\theta)$

• The gamma distribution approximates a normal distribution when $\kappa > 15$

teachingApps documentation built on July 1, 2020, 5:58 p.m.