Training set-based performance measures for neural net hypothesis testing

Performance measures for neural network hypothesis testing are derived based on the statistics of the training set. The training set-based measures are contrasted with maximum a posteriori probability (MAP) test measures. It is shown that the training set-based and MAP test probabilities are equal if the training set is proportioned according to the prior probabilities of the hypotheses. Applications of training set-based measures are suggested for neural net and training set design.<>