Venue: The University of Sydney, Carslaw Seminar Room 535, level 5, Carslaw Building (http://www.maths.usyd.edu.au/u/About/Carslaw.html)
Date: Aug 25th, 2015
Time: 6:00pm – 6.30pm: Refreshments
6:30pm – 7.30pm: Lecture
7:30pm – 8:00pm: Dinner (at a nearby restaurant)
Interactive and data adaptive model selection with mplot
This talk focuses on the computational aspects of selection criteria that are based on either inclusion or exclusion frequencies. We have developed the mplot R package which provides a collection of functions to aid exploratory variable selection. The package contains fast routines to make available modified versions of the simplified adaptive fence procedure (Jiang et al., 2009, Annals of Statistics) as well as other graphical tools such as variable inclusion plots and model selection curves (Mueller and Welsh, 2010, International Statistical Review; Murray et al, 2013, Statistics in Medicine). A browser based graphical user interface is provided to facilitate interaction with the results. These variable selection methods rely heavily on bootstrap resampling techniques. Fast performance for standard linear models is achieved using the branch and bound algorithm provided by the leaps package. The graphical model selection methods in mplot visualise popular model selection criteria that involve minimizing a penalized function of the data over a typically very large set of models. The penalty in the criterion function is controlled by a tuning parameter which determines the properties of the procedure. The implemented methods in mplot allow us to better explore the stability of model selection criteria through model selection curves and this is demonstrated through case studies.
Joint work with Garth Tarr (ANU); AH Welsh (ANU)
Keywords: Model selection; graphical methods; using R; information criterion; model selection curves; Fence
Samuel Mueller was born and educated in Switzerland and received his PhD in Mathematics from the University of Bern in November 2002. He joined the University of Sydney in 2008 as a Lecturer, was promoted to Senior Lecturer in 2010 and to Associate Professor in 2015. Previous appointments include a postdoc at the ANU (2003-2004) as well as academic positions at the University of Bern (2004-06) and the University of Western Australia (2006-2008). He served recently as the Postgraduate Director in the School of Mathematics and Statistics
(2012-2015) and is currently enjoying a sabbatical with various trips overseas and interstate. Samuel is a member of the Statistics Research Group. His specialties include Model Selection, Applied Statistics, Robust Methods and Extreme Value Theory. His research is mostly motivated by statistical problems that enable to learn more from omics, neuroscience and other complex data.