Introduction to Multiple Linear Regression
What you'll learn: How to extend simple linear regression from one predictor to many, modeling a response variable as a linear combination of multiple input features.
From One to Many Predictors
You've already learned simple linear regression, where you predict a response variable y using a single predictor x with the model ŷ = β₀ + β₁x. But what if you want to predict house prices using both square footage and number of bedrooms? Or model student performance using study hours and prior test scores?
Multiple linear regression extends the simple model to handle multiple predictors simultaneously.
The General Form
Instead of one predictor, you now have k predictors: x₁, x₂, ..., xₖ. The multiple linear regression model is:
ŷ = β₀ + β₁x₁ + β₂x₂ + ... + βₖxₖ
Where:
- ŷ is the predicted value of your response variable
- β₀ is the intercept (the baseline prediction when all predictors equal zero)
- β₁, β₂, ..., βₖ are the regression coefficients for each predictor
- x₁, x₂, ..., xₖ are your predictor variables
Each coefficient βᵢ represents the change in ŷ for a one-unit increase in xᵢ, holding all other predictors constant. This "holding constant" aspect is crucial—it's what distinguishes multiple regression from simply running several separate simple regressions.
An Analogy
Think of predicting crop yield. Simple regression uses only rainfall. Multiple regression uses rainfall, temperature, soil quality, and fertilizer amount—all simultaneously contributing to the prediction, each with its own weight (coefficient).
Key Takeaway: Multiple linear regression extends simple linear regression to multiple predictors using ŷ = β₀ + β₁x₁ + β₂x₂ + ... + βₖxₖ, where each coefficient represents the effect of one predictor while holding all others constant.