Random Variables

tags

Statistics

Introduction

It frequently occurs that in performing an experiment, we are mainly interested in some function of the outcome rather than the outcome itself. These quantities of interest, are real-valued functions defined on the sample space $\mathcal{S}$, known as random variables.

Discrete Random Variables

For a discrete random variable, we define the probability mass function $p(a)$ of $X$ by:

$$p(a)=P(X=a)$$

The probability mass function $p(a)$ is positive for at most a countable number of values of $a$.

Since $X$ must take on one of the values $x_i$, we have $\sum _{i=1}^{\mathrm{\infty }}p(x_i)=1$.

The cumulative distribution function $F$ can be expressed in terms of $p(a)$ by:

$$F(a)=\sum _{\text{all }{x}_i\le a}p(x_i)$$

Bernoulli Random Variable

A random variable $X$ is said to be a Bernoulli random variable if its probability mass function is given by:

$$p(0)=P(X=0)=1-p,p(1)=P(X=1)=p$$

This corresponds to the outcome of a trial with binary outcomes.

Binomial Random Variable

Suppose $n$ independent trials, each of which have binary outcomes and result in a success with probability $p$. If $X$ represents the number of successes that occur in the $n$ trials, then $X$ is a binomial random variable denoted by $Bin(n,p)$.

The pmf of a binomial random variable is given by:

$$p(i)=P(X=i)=(\genfrac{}{}{0ex}{}{n}{i})p^i(1-p{)}^i,i=0,1,\dots ,n$$

Geometric Random Variable

Suppose that independent trials, each having probability $p$ of being a success, are performed until a success occurs. If we let $X$ be the number of trials required until the first success, then $X$ is said to be a geometric variable with parameter $p$.

$$p(n)=P(X=n)=(1-p{)}^{n-1}p$$

Poisson Random Variable

A random variable $X$, taking on one of the values $0,1,2,\dots$ is said to be a poisson random variable with parameter $\lambda$ if for some $\lambda >0$,

$$p(i)=p(X=i)=e^{-\lambda }\frac{{\lambda }^i}{i!},i=0,1,\dots$$

An important property of the Poisson random variable is that it may be used to approximate a binomial random variable when the binomial parameter $n$ is large and $p$ is small.

Continuous Random Variables

Continuous random variables have an uncountable set of possible values.

$$P[X\in B]={\int }_{B}f(x)dx$$

The cumulative distribution $F(\cdot )$ and the probability density function is expressed by:

$$F(a)=P(X\in (-\mathrm{\infty },a])={\int }_{-\mathrm{\infty }}^{a}f(x)dx$$

Uniform Random Variable

$X$ is a uniform random variable on the interval $(a,b)$ if its probability density function is given by:

$$f(x)=\{\begin{array}{ll}\frac{1}{b-a}& a<x<b\\ 0& \text{otherwise}\end{array}$$

Exponential Random Variables

$$f(x)=\{\begin{array}{ll}\lambda e^{-\lambda x}& x\ge 0\\ 0& x<0\end{array}$$

$$F(a)=1-e^{-\lambda },a\ge 0$$

Gamma Random Variables

$$f(x)=\{\begin{array}{ll}\frac{{\lambda }^{\alpha }}{\mathrm{\Gamma }(\alpha )}{x}^{\alpha -1}{e}^{-\lambda x}& x\ge 0\\ 0& x<0\end{array}$$

The quantity $\mathrm{\Gamma }(\alpha )$ is called the gamma function and is defined by:

$$\mathrm{\Gamma }(\alpha )={\int }_0^{\mathrm{\infty }}{e}^{-x}{x}^{\alpha -1}dx$$

and $\mathrm{\Gamma }(n)=(n-1)!$ for some integral $\alpha =n$.

Normal Random Variables

$$f(x)=\frac{1}{\sqrt{2\pi }\sigma }{e}^{-\frac{(x-\mu {)}^2}{2{\sigma }^2}},-\mathrm{\infty }<x<\mathrm{\infty }$$

if $XN(\mu ,{\sigma }^2)$ and $Y=aX+b$, then $YN(a\mu +b,a^2{\sigma }^2)$.

Expectation of Random Variables

The expected value of a discrete random variable $X$ is defined by:

$$E(X)=\sum _{x:p(x)>0}xp(x)$$

and for a continuous random variable is defined similarly:

$$E(X)={\int }_{-\mathrm{\infty }}^{\mathrm{\infty }}xf(x)dx$$

Suppose we are interested in getting the expected value of a function of a random variable $E(g(X))$. One way is to obtain the distribution of $g(X)$, and compute $E(g(X))$ by definition of expectation.

$$E(g(X))=\sum _{x:p(x)>0}g(x)p(x)$$

We also have linearity of expectations:

$$E(aX+b)=aE(X)+b$$

$E(X)$ is the first moment of $X$, and $E(X^n)$ is the nth moment of $X$.

The variance of $X$, $Var(X)$ measures the expected square of the deviation of $X$ from its expected value:

$$Var(X)=E((X-E(X){)}^2)=E(X^2)-(E(X){)}^2$$

Jointly Distributed Random Variables

The joint cumulative probability distribution function of $X$ and $Y$ is given by:

$$F(a,b)=P(X\le a,Y\le b),-\mathrm{\infty }<a,b<\mathrm{\infty }$$

The distribution of $X$ can then be obtained from the joint distribution of $X$ and $Y$ as follows:

$$\begin{array}{rl}{F}_X(a)& =P(X\le a,Y\le \mathrm{\infty })\\ & =F(a,\mathrm{\infty })\end{array}$$

Covariance and Variance of Sums of Random Variables

The covariance of any 2 random variables $X$ and $Y$, denoted by $Cov(X,Y)$, is defined by:

$$\begin{array}{rl}Cov(X,Y)& =E[(X-E[X])(Y-E[Y])]\\ & =E[XY]-E[X]E[Y]\end{array}$$

If $X$ and $Y$ are independent, then $Cov(X,Y)=0$.

Some properties of covariance:

1. $Cov(X,X)=Var(X)$
2. $Cov(X,Y)=Cov(Y,X)$
3. $Cov(cX,Y)=cCov(X,Y)$
4. $Cov(X,Y+Z)=Cov(X,Y)+Cov(X,Z)$