First Ideas of Pattern Recognition

In scientific work the problem of determining or predicting an obscure property of an object or an event often arises. The property is called obscure because only indirect measurements can be made that are known or supposed to be related to the property in question. If a theoretical relationship between the measurements and the property is not completely known a method of data interpretation may be used which is sometimes called educated guess C1533. In this approach an empirical relationship is derived from a collection of objects for which the interesting property and the indirect measurements are known. This relationship is then used to predict the obscure property of unknown objects. 

A set of indirect measurements which describe one object is called a pattern. The determination of the obscure property is often a recognition of the class (category) to which a pattern belongs. Classification of patterns is a fundamental process in many parts of science and human being and therefore mathematical methods of pattern recognition find wide applications in very different fields. 

An example from chemistry may characterize a typical pattern recognition problem The objects may be chemical compounds and the property to be determined is the presence of a carbonyl group in the molecule. 

---