Skip to content

new paper: how subclasses affect classification​

  • paper

Our new paper “Multivariate classification of neuroimaging data with nested subclasses: Biased accuracy and implications for hypothesis testing” was just published in PLOS Computational Biology.
In this paper, we show how subclasses can bias classification results and inflate correct classification rates of linear classifiers. We show, when this bias is strongest and how to use permutation tests to correct this bias.