Classification procedure

from Wikipedia, the free encyclopedia

Classification methods , also classification methods , are methods and criteria for dividing (classifying) objects or situations into classes, that is, for classification . Such a method is also known as a classifier . Many methods can be implemented as an algorithm ; one also speaks of machine or automatic classification. Classification methods are always application-related, so that there are many different methods.

In a narrow sense, in contrast to the classification methods, the classification methods are used to classify objects in existing classes. In colloquial terms, however, no distinction is made between classifying and classifying .

Classification processes play a role in pattern recognition , artificial intelligence , documentation science and information retrieval , among other things . Various parameters can be determined to assess a classifier .

Types of classification methods

Since a strictly hierarchical classification of classification procedures is hardly possible, they can best be classified using different properties:

  • Manual and automatic procedures
  • Numerical and non-numerical methods
  • Statistical and non-distribution methods
  • Supervised and unsupervised procedures
  • Firmly dimensioned and learning processes
  • Parametric and non-parametric methods

Manual and automatic procedures

With automatic methods, the classification takes place by means of an automatic process by software . The process of machine classification can be described as a formal method of making decisions in new situations based on learned structures . The automatic classification is a branch of machine learning .

More precisely, this is the generation of an algorithm (the learning algorithm) which - applied to known and already classified cases (the database) - calculates structures. These newly learned structures enable a further algorithm (the evaluating algorithm) to assign a new and previously unknown case to one of the known target classes on the basis of the observed attributes and their characteristics.

Statistical and non-distribution methods

Statistical methods are based on density calculations and probabilities, while non-distribution methods use clear dividing surfaces to separate the classes. The boundaries between the individual classes in the feature space can be specified using a discriminant function.

Examples of statistical methods are the Bayesian classifier , the fuzzy pattern classifier and the kernel density estimator . The calculation of parting surfaces is possible using so-called support vector machines .

Supervised and unsupervised procedures

The creation of structures from existing data is also known as pattern recognition , discrimination or supervised learning . Classifications are given, which can also be done by taking samples. In contrast to this, there is unsupervised learning , in which the classes of the data are not specified, but these must also be learned. However, this case can the reinforcement learning ( English reinforcement learning ) Information added about whether a classification was right or wrong. An example of unsupervised procedures is cluster analysis .

Parametric and non-parametric methods

Parametric methods are based on parametric probability densities , while non-parametric methods (e.g. closest neighbor classification ) are based on local density calculations.


See also


  • O. Oberhauser: Automatic classification. Development status - methodology - areas of application. Peter Lang, Frankfurt / Main et al. 2005. ISBN 3-631-53684-4 .
  • Andrew R. Webb : Statistical Pattern Recognition . John Wiley & Sons, 2nd ed., July 2002.
  • Richard O. Duda , PE Hart, DG Stork: Pattern Classification. John Wiley & Sons, 2nd ed., 2000.
  • CM Bishop: Neural Networks for Pattern Recognition. Oxford University Press, 1996.
  • D. Michie, DJ Spiegelhalter, CC Taylor (Eds.): Machine Learning , Neural and Statistical Classification. Elis Horwood, 1994.

Individual evidence

  1. HardWin Jungclaussen: Causal computer science. Introduction to the teaching of active linguistic modeling by humans and computers. Springer Fachmedien Wiesbaden , 2013, ISBN 978-3-322-81220-9 , p. 57.