Learning phase transitions: comparing PCA and SVM

Abstract

A comparison of results from principal component analysis and support vector machine calculations is made for a variety of phase transitions in two-dimensional classical spin models.

Comments: 22pages, 24 figures