June, 11-15, 2010
|Last update: 7/6/2010|
The Applied Bayesian Statistics summer school has been organised since 2004 byThe 2010 School will be organised in cooperation with the European Academy Bozen/Bolzano (EURAC), through the Institute of Genetic Medicine, whose fast growing statistics and bioinformatics groups are developing methods and tools for the analysis of large scale genomic data. IMATI acknowledges the support by the Information and Communication Technology (ICT) Department of the National Research Council (CNR).
The school aims to present state-of-the-art Bayesian applications, inviting leading experts in their field. Each year a different topic is chosen. Past editions were devoted to Gene Expression Genomics, Decision Modelling in Health Care, Spatial Data in Environmental and Health Sciences, Bayesian Methods and Econometrics, Bayesian Decision Problems in Biostatistics and Clinical Trials and Bayesian Methodology for Clustering, Classification and Categorical Data Analysis.
The topic chosen for the 2010 school is
Bayesian Machine Learning with Biomedical Applications.
The lecturer will be David Dunson, Department of Statistical Science, Duke University, Durham, NC, USA
He will be assisted by Raffaele Argiento (CNR IMATI, Italy).
This short course is intended to provide a practically-motivated introduction to Bayesian methods for machine learning and high-dimensional data analysis. Some topics of particular interest include high-dimensional variable selection for regression and classification, and multi-task learning and combining of information for related signals, functions or images. A brief overview will be provided of Bayesian methods for linear regression with very many predictors using shrinkage priors and mixture priors. This overview will include Bayesian formulations of Lasso, elastic net and relevance vector machine (RVM) methods that have been widely used in the literature. In addition, spike and slab mixture priors for formal Bayes subset selection and model averaging will be presented. We also describe sparse Bayesian latent factor regression methods, which can accommodate selection of correlated sets of predictors in large p, small n settings. Computational methods will be described based on maximum a posteriori estimation and Markov chain Monte Carlo algorithms. The methods will be compared through simulation studies and applied to a variety of data examples, including machine learning data and biomedical applications involving gene expression and other high-dimensional markers. An emphasis will be on practical issues in implementing and interpreting results, and code will be provided in R.
The school will make use of lectures, practical sessions, software demonstrations, informal discussion sessions and presentations of research projects by school participants. The slides and background reading material will be distributed to the students before the start of the course.
The prerequisites for this course will include basic knowledge of linear algebra, statistical inference, linear regression models, and data analysis. This course will be beneficial to graduate students, post-docs and researchers both from academia and industry whose area of activity is Statistics, Mathematics, Actuarial Science, Computer Science, Biostatistics and Genomics, Epidemiology, and Engineering. Some knowledge of Bayesian Statistics is desirable, but not required. Prospective students may refer to the book Bayesian data analysis, 2nd Ed. by Andrew Gelman, John B. Carlin, Hal S. Stern and Donald B. Rubin (Chapman & Hall/CRC Texts in Statistical Science Series, 2004).
The 2010 school will be held at European Academy (EURAC), located in the city of Bolzano (Bozen in German), in the Italian Alps, not far from the Austrian border and the Dolomites.
Please note that the number of available places is limited.
The school will start on Friday, June, 11th, at 9.00 and it will end on Tuesday, June, 15th, at 13.00. Sunday afternoon (June, 13th) is free. Welcome buffet and farewell dinner are planned on June, 11th and 14th, respectively.
The registration fees are: