« Back to Glossary Index

In Machine Learning, regression is a technique to understand the relationship between independent variables or features and a dependent numerical variable, or outcome, with the main purpose of estimating / predicting the outcome. It is computed fitting a function, usually a line or a polynomial to datapoints in order to minimise an error function. There are several types of regression based on the fitting and error functions: for instance, Linear Regression, Polynomial Regression, Bayesian Regression, Ridge Regression, LASSO Regression, etc.