Title<\/strong>:\u00a0\u00a0 Regularized\u00a0 version of the minimax\u00a0 probability machine<\/p>\n Abstract:<\/b> In this talk we present novel second-order cone programming formulations for binary classification, by extending the Minimax Probability Machine (MPM) approach. Inspired by Support Vector Machines, a regularization term is included in the MPM and Minimum Error Minimax Probability Machine methods. This inclusion reduces the risk of obtaining ill-posed estimators, stabilizing the problem, and, therefore, improving the generalization performance. Our approaches are first derived as linear methods, and subsequently extended as kernel-based strategies for nonlinear classification. Experiments on well-known binary classification datasets demonstrate the virtues of the regularized formulations in terms of predictive performance.<\/p>\n A recorded video of the conference is …. ;\u00a0 the slides can be downloaded\u00a0 …<\/p>\n Venue<\/strong>: Online via Google Meet: https:\/\/meet.google.com\/gjq-cwic-udo<\/span><\/span><\/span><\/p>\n A brief biography of the speaker<\/span><\/span><\/strong>: Julio L\u00f3pez is an Associate Professor at the Institute of Basic Sciences, University Diego Portales, Santiago, Chile. He obtained his PhD degree at the University of Chile, in 2009. <\/span><\/span>His research interests include conic programming, convex analysis, algorithms, and machine learning.<\/span><\/span><\/p>\n Coordinators<\/strong>: Fabi\u00e1n Flores-Baz\u00e1n (CMM, Universidad de Concepci\u00f3n) and Abderrahim Hantoute (Alicante)<\/p>\n","protected":false},"excerpt":{"rendered":" Speaker: Professor Julio L\u00f3pez Instituto de Ciencias B\u00e1sicas, Universidad Diego Portales, Santiago, Chile Date:\u00a0 September 08,\u00a0 2021 at 09:30 am (Chilean-time) Title:\u00a0\u00a0 Regularized\u00a0 version of the minimax\u00a0 probability machine Abstract: In this talk we present novel second-order cone programming formulations for binary classification, by extending the Minimax Probability Machine (MPM) approach. Inspired by Support Vector … Continue reading Regularized version of the minimax probability machine<\/span>