For questions related to the placement of individual cases into categories, such as is essential in fraud detection, spam detection, quality control, prediction of user or market responses, automated organizing or indexing, assigning objects in view to types of obstacles or risks, writing or typing recognition, phonic recognition, .

Classification is the automatic categorization of a new observation. This classification is based on a model produced from a training set of data containing observations whose classifications are given. Classification is especially useful for problems involving categorical data.