Supervised learning is a statistical framework in which we use training data comprising predictors (independent variables, features, X) and an outcome (dependent variable, target, y) to train a model that can predict y given some X. In other words, we are trying to learn the patterns of X that relate to y.
This contrasts unsupervised learning, where there is no outcome we are trying to predict, and we’re instead concerned with discovering patterns in how various features of X relate to one another.