Learning theory, linear seperability and noisy data
Griff L. Bilbro, Wesley E. Snyder · NCSU Libraries Repository (North Carolina State University Libraries) · 1990
We apply the statistical mechanical theory of learning [6] to the determination from examples of a line separating two classes, a simple enough example to admit intuition and comparision.We report the resulting dependence of the average prediction probability on the size, quality, and other features of the training set as well as the error to which the system is trained on the set. 1