Regression is primarily used when we are trying to model continuous numerical response variable. Classification is used when we are trying to predict categorical response variable.
Examples:
- Estimating how much a customer will spend in the upcoming month is Regression.
- Predicting if the customer will default or not is Classification