A machine learning approach where algorithms learn from labeled training data, with input-output pairs provided during training. The model learns to map inputs to correct outputs, enabling predictions on new, unseen data for tasks like classification and regression problems.
