Skip to content

What is Regression?

Regression is a supervised machine learning method that predicts continuous numerical values from input features.

In regression, a model learns the relationship between input variables and a target output that can take any real number value, such as price or temperature. The training process adjusts model parameters to minimize the difference between predicted and actual values, often using a loss function like mean squared error.

Common algorithms include linear regression, which fits a straight line, and more flexible methods like decision trees or neural networks. Key considerations involve checking assumptions, handling overfitting through regularization, and evaluating performance with metrics such as R-squared or mean absolute error.

Unlike classification, regression outputs are not limited to discrete categories, making it suitable for forecasting and trend analysis tasks.

Example

A real-estate app uses regression to estimate a house's selling price from features like square footage, number of bedrooms, and location, outputting a dollar amount rather than a category.

Why it matters

Regression powers everyday predictions in finance, healthcare, and demand forecasting, forming a core building block for many modern AI systems that need numeric outputs.

Frequently asked questions

Regression predicts continuous numbers while classification assigns data to discrete categories.