A Sequential Decision Model for Optimum Recognition

It has been recognized in recent years that the problem of pattern recognition can be viewed as a test of statistical hypotheses. A concise explanation of this theoretical approach to the recognition problem may be obtained by considering the statistical decision process as a zero-sum two-person statistical game. The two players will be referred to as nature and the decision maker, i.e., the recognition system. A non-negative weighting function measures the loss incurred by the decision maker as a consequence of each possible decision. The aim of the decision maker is to minimize the expectation of the loss in recognition.