Supervised learning
Supervised learning is a machine learning process that teaches algorithms predictive models by feeding the algorithm input data with known outcomes—i.e., supervised learning teaches algorithms by example.
The input/output pair (labelled data) provides feedback for the algorithm, which uses the training data set to adjust the model to minimise error.
For example, a training set may contain pictures of different kinds of animals with a label associated to each picture, allowing the algorithm to compare the predicted label with the correct one.
Supervised learning uses a validation data set to measure the algorithm’s progress in learning the model and a test data set to evaluate the model’s performance on never-before-seen data to determine whether the model has learned its training data effectively and can generalise to new data.
Sources: