foundations of computational agents
A learned model is a function from the input features to the target features. Most supervised learning methods take the input features, the target features, and the training data and return a compact representation of a function that can be used for future prediction. An alternative to this is case-based reasoning, which uses the examples directly rather than building a model. Learning methods differ in which representations are considered for representing the function. This section considers some basic models from which other composite models are built. Section 7.6 considers more sophisticated models that are built from these basic models.