Learning Curves for Error Minimum and Maximum Likelihood Algorithms

For the problem of dividing the space originally pad tionec ~y a blurred boundary, every learning algorithm can make the probability of incorrect prediction of an individual example E decrease with the number of training examples t. We address here the question of how the asymptotic form of ~ ( t ) as well as its limit of convergence reflect the choice of learning algorithms. The error minimum algorithm is found to exhibit rather slow convergence of E ( t ) to its lower bound