More on Random Variables

A random variable is a value which we do not know yet or we are uncertain of. They are the fundamental unit of statistics and are used anytime we want to predict a value. Having a good understanding of their properties makes it easier to model situations with statistics. In general, we know the values that a random variable could take on, and for each value, the probability of that value occurring.

Let’s take the canonical example: a die roll. We will use a random variable DD to represent the value after a single roll of a fair die. It's important to understand that all random variables have a story or context behind them. In this scenario, when we roll a dice, there are 6 possible values (1, 2, ..., 6), each equally likely of occurring (a probability of 1/6).

This is how random variables are usually explained. However, more recently, I learned some more definitions that I found simple and easy to understand.

The first being BB-valued random variables.

Given a set BB, we say that the random variable XX is BB-valued if P[xB]=1P[x \in B] = 1.

In other words, the probability that a BB-valued random variable takes on a value that is in the set BB is 1. Using our example above, the random variable, DD, representing a die roll is {1,2,3,...,6}\{1, 2, 3, ..., 6\}-valued. It also follows that the DD is also {1,2,...,6,7}\{1, 2, ..., 6, 7\}-valued; the set BB does not necessarily need to be the smallest set that contains the values of the random variable. Essentially, as long as BB contains the values 1...61...6, then DD is BB-valued. Therefore, we can say DD is also N\mathbb{N} and R\mathbb{R}-valued.

Just knowing a set BB is usually not sufficient to describe a random variable since it may contain many values that our random variable will never take. The next definition will help.

The support SX\mathcal{S}_X of the discrete random variable XX is the smallest set such that XX is SX\mathcal{S}_X-valued.

So, while our random variable DD might be N\mathbb{N}-valued, its support SD\mathcal{S}_D is {1,2,...,6}\{1, 2, ..., 6\}. This is the smallest set BB such that P[xB]P[x \in B] is 1; in otherwords, if we were to remove any value from SD\mathcal{S}_D and form BB, then the P[xB]<1P[x \in B] < 1.

Now we can use these two concepts to discuss other concepts we already know. For example, we may use these definitions to formalize the definition of discrete random variables. Or maybe characterize types of random variables such as Bernoulli random variables.

WED JAN 08 2020
Website built on top of the work of .
SCROLL: 0.00#808080