Simple properties about gamma distribution and poisson distribution
gamma distribution
p.d.f
$\Gamma(\alpha,\beta)=\frac{1}{\Gamma(\alpha)}x^{\alpha-1}\beta^{\alpha}e^{-\beta x}$ $x>0$
k-th moment
$$
\begin{equation}
E[X^k]=\int_0^{\infty} \frac{1}{\Gamma(\alpha)}x^{\alpha-1}x^{k}\beta^{\alpha}e^{-\beta x}=\frac{\Gamma(k+\alpha)}{\Gamma(\alpha)}\frac{1}{\beta^k}
\end{equation}
$$
So, we can get $E[X]=\frac{\alpha}{\beta}$ and $Var(X)=\frac{\alpha}{\beta^2}$
moment generating function
$\left( 1-\frac{t}{\beta}\right)^{-\alpha}$ $t<\beta$
characteristic function
$\left( 1-\frac{it}{\beta}\right)^{-\alpha}$
Using the properties of characteristic we can get the properties as follows
properties
- If $X_1 \sim \Gamma(\alpha_1,\beta)$ and $X_2 \sim \Gamma(\alpha_2, \beta)$, then $X_1+X_2 \sim \Gamma(\alpha_1+\alpha_2,\beta)$
- If $X \sim \Gamma(\alpha,\beta)$, then $kX \sim \Gamma(\alpha,\frac{\beta}{k})$
- Exponential distribution $Exp(\lambda)\sim \Gamma(1,\lambda)$ and Chi square distribution $\chi^2(n)\sim \Gamma(\frac{n}{2},\frac12)$
poisson distribution
p.m.f (probability mass function)
$P(X=x)=\frac{\lambda^x}{x!}e^{-\lambda x}$ $x\in \mathbb{N}$
characteristic function
$X\sim P(\lambda)$
$$
\begin{equation}
E[e^{itX}]=\sum_{k=0}^{\infty}e^{itx}\frac{\lambda^x}{x!}e^{-\lambda x}=e^{-\lambda}e^{\lambda e^{it}}\sum_{k=0}^{\infty}
\frac{(e^it\lambda)^x}{x!}e^{-\lambda e^it}=e^{\lambda(e^{it}-1)}
\end{equation}
$$
Using the properties of characteristic we can get the property as follows
property
If $ X_1, X_2 \dots X_n $ i.i.d $\sim P(\lambda)$, then $\sum_{i=1}^{n}X_i \sim P(n\lambda)$