概率论常用定理

---

Central Limit Theorem （ 中心极限定理）

Independent Identically Distribution （独立同分布的中心极限定理 ）

Definition

It refers that a sequence of random variables and its distribution can be asymptotically approximated to Normal/Gaussian distribution.

Let random variables $X1，X2，......X_n$ be Independently Identical distributed(i.i.d), and we have $\mathbb{E}(X_i)=\mu ,\mathbb{D}(X_i)={\sigma }^2,(i=1,2...,n)$. For any $x$, the distribution function
$$F_n(x)=P\left\{\frac{{\sum }_{i=1}^n{X}_i-n\mu }{\sigma \sqrt{n}}\le x\right\}$$ satisfies
$$\underset{n\to \infty }{\lim }{F}_n(x)=\underset{n\to \infty }{\lim }\left\{\frac{{\sum }_{i=1}^n{X}_i-n\mu }{\sqrt{n}\sigma }\le x\right\}=\frac{1}{\sqrt{2\pi }}{\int }_{-\infty }^x{e}^{-\frac{-t^2}{2}}dt=\varnothing (x).$$
This theorem give the fact: for enough $n$, random variable $Y_n=\frac{{\sum }_{i=1}^n{X}_i-n\mu }{\sqrt{n}\sigma }$ can be approximated to standard normal distribution, i.e., $Y_n\sim N(0,1)$. Thus, for enough $n$, ${\sum }_{i=1}^n{X}_i=\sqrt{n}\sigma Y_n+n\mu$ can be approximated to normal distribution $N(n\mu ,n\sigma )$.
Note that for sufficiently large $n$, the sum of $n$ Independently Identically Distributed random variables can be considered as a normal variable.

Different Distributed Central Limit Theorem

Let $X_1,X_2,...,X_n$ and $f_{x_k}(x)$ for $k=1,...,n$ be a list of independent random variables and probability density function, respectively, with the following mean and variable:
$$\mathbb{E}(X_k)=\mu k,\mathbb{D}(X_k)={\sigma }_k^2.$$ Then, we define $$B_n^2=\sum _{i=1}^n{\sigma }_k^2$$ and $$Y_n=\frac{{\sum }_{k=1}^n{X}_k-{\sum }_{k=1}^n\mu k}{B_n}.$$

For any positive number $\tau$,
$$\underset{n\to \infty }{\lim }\frac{1}{B_n^2}\sum _{k=1}^n{\int }_{|x-{\mu }_k>\tau B_n|}{\left(x-{\mu }_k\right)}^2{f}_{x_k}(x)dx=0.$$

For any $x$, the distribution function $F_n(x)$ of random variable $Y_n$ should satisfy $$\underset{n\to \infty }{\lim }{F}_n(x)=\underset{n\to \infty }{\lim }P\left\{\frac{{\sum }_{k=1}^n{X}_k-{\sum }_{k=1}^n\mu k}{B_n}\le x\right\}=\frac{1}{\sqrt{2\pi }}{\int }_{-\infty }^x{e}^{-\frac{t^2}{2}}\text{d}t.$$

This theorem proves that the distribution function can be approximated to normal/Gaussian distribution, if the random variables have a large number of independent and uniform random variables.

---