Pattern Recognition

From Chessprogramming wiki
Jump to: navigation, search

Home * Knowledge * Pattern Recognition

Chess pattern [1]

Pattern Recognition,
is used to assign a label to an input value [2] , for instance to apply classification in machine learning applications, i.e. to identify objects and images, as well as computer chess related pattern of chess positions in Cognitive Psychology and concerning evaluation and control of the search in computer chess. Pattern recognition algorithms generally aim to provide a reasonable answer for all possible inputs and to perform a "fuzzy" matching. In contrast, Pattern matching usually has to be exact.

Chess Pattern

Chess pattern range from simple properties of squares and pieces concerning occupancy and control, to a more complex interrelated sets of features. Recognizers are implemented with decision trees, neural networks, and fuzzy logics. In his ICCA Journal paper Fuzzy Production Rules in Chess, Peter W. Frey [3] proposed feature strings or sets of three types. Type-A features must match completely, type-B feature strings represent features which are usually but not always present, while type-C features are present occasionally but are highly diagnostic when available. Those features were intended to use at the root for an oracle approach.

Chess Programs

See also

Publications

1950 ...

1960 ...

1970 ...

1980 ...

Revised as Hans Berliner, Carl Ebeling (1990). Hitech. Computers, Chess, and Cognition

1990 ...

2000 ...

2010 ...

2020 ...

Forum Posts

External Links

Hans Ulrik, Makiko Hirabayashi, Klavs Hovman, Marilyn Mazur

References

Up one Level