2 Minimum Classification Error Criterion for Recognizer Design : The practical issues discussed above against the theoretical optimality of the MAP- based classifier design motivate the search for alternative designs of classifiers. In this section, we discuss one such alternative, namely the use of the minimum classification error (MCE) criterion. This particular approach was originally proposed in [Juang 92],

Chapter

Chapter

ABSTRACT

Given a discriminant function, the misclassification measure for the l-th input training observation sample xt from class IS is defined as

Loss function

The loss function is defined as a sigmoid, non-decreasing function of d,,:

which approximates the classification error count.