Statistical Mechanics of Learning in the Presence of Outliers
arXiv:cond-mat/9803316 · doi:10.1088/0305-4470/31/46/005
Abstract
Using methods of statistical mechanics, we analyse the effect of outliers on the supervised learning of a classification problem. The learning strategy aims at selecting informative examples and discarding outliers. We compare two algorithms which perform the selection either in a soft or a hard way. When the fraction of outliers grows large, the estimation errors undergo a first order phase transition.
24 pages, 7 figures (minor extensions added)