International Society for Bayesian Analysis (ISBA) The main ensemble methods discussed are Bayes model averaging, bagging, and stacking. There are 5 sections: 1) General overview 2) Neural nets, 3) Trees, 4) Ensemble method, and 5) Applications with real data. At several points in the development standard test, functions are also used to compare methods. ISBA 2012 PowerPoint Slides June 25-29, 2012 | Kyoto, Japan ...
Read MoreClustering Stability: Impossibility and Possibility This presentation was addressed to a specialized audience of people in Data Mining and Machine Learning. The talk provided a theory that showed how clustering stability can be used to choose the correct number of clusters, as well as demonstrate the importance of cluster stability and discussed the use of stability for clustering evaluation. Bertrand Salem Clarke (UM), Hoyt Koepke...
Read MoreClustering Impossibility and Stability This presentation was aimed at a general statistical audience and showed the importance of cluster stability (a theorem that in the limit of large dimensions can get complete noninformativity of clustering for finite sample sizes) and discussed the use of stability for clustering evaluation. Bertrand Salem Clarke (UM), Hoyt Koepke (UM and U. Washington) Department of Statistics, Florida State...
Read MoreDimension Reduction (Mostly) This presentation was a review of the basics of machine learning with commentary on the major dimension reduction techniques. Bertrand Salem Clarke (UM) Notre Dame University Conference Louaize, Lebanon | July 4-9, 2011 NDU Powerpoint slides...
Read MoreReview of Bayesian and Frequentist Statistics This presentation was part 1 of a 4-part course given on the basics of machine learning. This first course was a review of standard graduate statistics material. Bertrand Salem Clarke (UM) Bertrand Salem Clarke (UM)Notre Dame University Conference Louaize, Lebanon | July 4-9, 2011 NDU 2011-Pt.1 PowerPoint Slides 4-9...
Read More