Random variables are the building blocks of statistical modeling. They help us quantify uncertain outcomes in experiments or real-world scenarios. By assigning numerical values to events, we can analyze and predict patterns in data.
Understanding the types and properties of random variables is crucial. Whether discrete or continuous, univariate or multivariate, these concepts form the foundation for probability distributions and statistical inference techniques we'll explore later.
Understanding Random Variables
Discrete vs continuous random variables
- Discrete random variables take on countable, distinct values limited to specific numbers (number of customers in a store, heads in coin tosses, defective items in a batch)
- Continuous random variables assume any value within a continuous range measuring quantities on a scale (height of a person, time until next bus arrival, temperature in a room)
Support and range of variables
- Support encompasses all possible values a random variable can take defining its domain
- For discrete variables, support lists individual values (die roll: {1, 2, 3, 4, 5, 6})
- For continuous variables, support specifies interval or union of intervals (height: [0, ∞))
- Range sets all possible outcomes often matching support but may differ in specific cases
Univariate vs multivariate variables
- Univariate random variables describe single characteristic represented by one variable (X = temperature in ℃)
- Multivariate random variables describe multiple characteristics simultaneously represented by vector of variables ((X, Y) = (temperature in ℃, humidity %))
Random variables as functions
- Function perspective maps outcomes from sample space to real numbers assigning numerical values to events
- Domain comprises sample space of experiment while codomain includes set of real numbers
- Formalization: Let S be sample space, X: S → R is random variable
- Enables mathematical operations on outcomes and probability calculations
- Mapping examples: Coin toss (Heads → 1, Tails → 0), Die roll (Face value → corresponding number)