2021-10-21 Laplace分布和指数分布，相同参数的两个指数分布之差是Laplace分布

先给出一些公式：

Laplace分布$Lap(b)$的probability density function:
$$f(x;\mu ,b)=\frac{1}{2b}\exp (-\frac{|x-\mu |}{b})=\frac{1}{2b}\{\begin{array}{rlr}& \exp (-\frac{\mu -x}{b}),& x<\mu \\ & \exp (\frac{x-\mu }{b}),& x\ge \mu \end{array}$$
cumulative distribution function:
$$F(x)=\{\begin{array}{rlr}& \frac{1}{2}\exp (\frac{x-\mu }{b}),& x<\mu \\ & 1-\frac{1}{2}\exp (\frac{\mu -x}{b}),& x\ge \mu \end{array}$$
指数分布$Expo(\lambda )$的probability density function:
$$f(x;\lambda )=\{\begin{array}{rlr}& \lambda e^{-\lambda x},& x\ge 0\\ & 0,& x<0\end{array}$$
cumulative distribution function:
$$F(x)=\{\begin{array}{rlr}& 1-e^{-\lambda x},& x\ge 0\\ & 0,& x<0\end{array}$$

Lemma.$Y=X_1-X_2$，且$X_1,X_2\sim Expo(\lambda )$，则$Y\sim Lap(\frac{1}{\lambda })$.【注意：$X_1,X_2$的$\lambda$不相等时，$Y$并不是Laplace分布】

Proof.
$$\begin{array}{r}{f}_Y(y)=f_{X_1-X_2}(y)={\int }_{x_1=0}^{\infty }{f}_{X_1}(x_1)f_{X_2}(x_1-y)d{x}_1\end{array}$$
考虑到$f_{X_2}$时分段函数，因此上述式子需要做分类讨论。当$y\ge 0$时，
$$\begin{array}{rl}{f}_Y(y)=& {\int }_{x_1=y}^{\infty }{f}_{X_1}(x_1)f_{X_2}(x_1-y)d{x}_1\\ =& {\int }_{x_1=y}^{\infty }\lambda e^{-\lambda x_1}\lambda e^{-\lambda (x_1-y)}d{x}_1\\ =& \lambda e^{\lambda y}{\int }_{x_1=y}^{\infty }\lambda e^{-2\lambda x_1}d{x}_1\\ =& \frac{\lambda }{2}{e}^{\lambda y}{e}^{-2\lambda y}\\ =& \frac{\lambda }{2}{e}^{-\lambda y}\end{array}$$
当$y<0$时，
$$\begin{array}{rl}{f}_Y(y)=& {\int }_{x_1=0}^{\infty }{f}_{X_1}(x_1)f_{X_2}(x_1-y)d{x}_1\\ =& {\int }_{x_1=0}^{\infty }\lambda e^{-\lambda x_1}\lambda e^{-\lambda (x_1-y)}d{x}_1\\ =& \lambda e^{\lambda y}{\int }_{x_1=0}^{\infty }\lambda e^{-2\lambda x_1}d{x}_1\\ =& \frac{\lambda }{2}{e}^{\lambda y}\end{array}$$

综上，得$f_Y(y)=\frac{\lambda }{2}{e}^{-\lambda |y|}$.