Supervised Learning

Supervised Learning is a fundamental paradigm in Machine Learning in which a Model learns from a labeled Dataset containing inputs paired with known outputs or targets. Its goal is to learn a Mapping Function that accurately predicts outputs for new, unseen inputs. Training commonly involves minimizing a Loss Function through Optimization methods such as Gradient Descent, but gradient descent is not required: other training procedures and, for some models, Closed-Form Solutions are available. Workflows often use a Training Set to fit the model and a Validation Set to guide model selection and assess Overfitting, sometimes with a separate Test Set to estimate Generalization. These arrangements vary; a validation set is not mandatory, and splitting data does not guarantee generalization or prevent overfitting. Common tasks include Classification, which predicts discrete categories, and Regression, which predicts continuous numerical values. Representative algorithms and model families include Support Vector Machines, Decision Trees, Linear Regression, and Neural Networks.