Output details
11 - Computer Science and Informatics
University of Birmingham
Classification of mislabelled microarrays using robust sparse logistic regression
<28>Paper appearing in a journal ranked #1, by impact factor, among journals on Mathematical & Computational Biology. It develops a label-robust sparse classifier for gene array data. The gain in performance demonstrates that dealing with label noise is likely to be a substantial element in advancing micro-array analysis. The core methodology is generic, and appears in Proc. ECML-PKDD(1), 2012, pp. 143-158 (acceptance rate 20%), with further advances in Proc. UAI 2013, pp. 82-90, 2013 – a top venue. These works led to Kaban being a co-guest editor of a forthcoming special issue of Neurocomputing devoted to “Learning with Label Noise”.