Pattern Recognition, Statistical
Thomas Frank Hofmann · Encyclopedia of Cognitive Science · 2005
Abstract Statistical pattern recognition deals with the problem of automatically classifying objects as belonging to one or more classes from a set of possible classes. Objects are typically represented by raw data that may include measured or known object properties and which are summarized in a feature vector. The general goal is to infer suitable classification rules that map feature vectors to class labels based on a training set of objects with known class memberships.