Recognition With Learning As a Problem of Choice

Abstract

The paper highlights a problem of pattern recognition with learning as a part of a wider problem of choice. It is shown that by information basis the problem can be compared with problems of formulae classification in sentence calculus and logical diagnostics. A family of algorithms is constructed that can be used for solving any of the above-mentioned problems. It is shown that an algorithm from the family can be compared with classical ones (method of resolution, etc), and, therefore, is justified at least in the algorithmic sense.