Classification and Regression
The methods used for a model depend on the type of output desired.
Classification is the attribution of a category (or class) to a given inputs. The set of possible classes is usually finite and static. For example, the possible classes may be the set of alphanumeric characters when attempting to recognise handwriting. The outputs from classification models are discrete.
Regression is the prediction of a specific value for a given input, using interpolation and extrapolation on the training data. A regression model make take many forms, such as linear or polynomial. The outputs from regression models are continuous.