Dear list members,
Greetings and apologies for cross-posting.
The following courses are scheduled to take place at the Statistical Services Centre in November 2016. Summary information is given below. For registration forms please see http://www.reading.ac.uk/ssc/, or email [log in to unmask]
******************************************************************************************************************
Course 1: Bayesian Modelling, Inference, Prediction and Decision-Making
Date: 21-22 November 2016.
Duration: 2 days.
View: http://www.reading.ac.uk/ssc/training/CourseDetails.php?name=Bayesian_Modelling,_Inference,_Prediction_and_Decision-Making
******************************************************************************************************************
Course 2: Bayesian Hierarchical Modelling
Date: 23 November 2016.
Duration: 1 day.
View: http://www.reading.ac.uk/ssc/training/CourseDetails.php?name=Bayesian_Hierarchical_Modelling
******************************************************************************************************************
Course 3: Bayesian Model Specification and Bayesian Non-Parametric Modelling
Date: 24 November 2016.
Duration: 1 day.
View: http://www.reading.ac.uk/ssc/training/CourseDetails.php?name=Bayesian_Model_Specification_and_Bayesian_Non-Parametric_Modelling
******************************************************************************************************************
Course 4: Case Studies in Bayesian Data Science in an Era of 'Big Data'
Date: 25 November 2016.
Duration: 1 day.
View: http://www.reading.ac.uk/ssc/training/CourseDetails.php?name=Case_Studies_in_Bayesian_Data_Science_in_an_Era_of_Big_Data
******************************************************************************************************************
Prices: 370 GBP for any one day; 695 GBP for any two days; 995 GBP for any three days; 1285 GBP for any four days; 1565 GBP for all 5 days.
A 30% academic discount is available for these courses. [Terms and conditions apply.] Note a special discount of 50% will be given to students for these specific courses. Please indicate you wish to apply for a discount when you register, together with information supporting your eligibility. [Terms and conditions apply.]
Please view the individual course web pages, given above, for more information. Further information is also given below.
**********************************************************************************************
Outline of week of courses
**********************
This is a set of short courses that offers you flexibility: you can either
take:-
* the first course (two days), which is an introductory course on Bayesian modelling, inference, prediction and decision-making
* or the second course (one day), which is an intermediate-level course on Bayesian hierarchical modelling,
* or the third course (one day), which is an intermediate-level course on Bayesian model specification,
* or the fourth course (one day), which aims to illustrate a variety of best-practice data science methods using datasets that are potentially enormous,
* or any combination of the courses listed above.
Course 1 will :
*********************
(1) Compare and contrast the frequentist and Bayesian conceptions of probability, highlighting the strengths and weaknesses of both;
(2) Review maximum-likelihood fitting of statistical models;
(3) Show you how to obtain Bayesian solutions to inferential and predictive problems analytically and in closed form (when such solutions are available); and
(4) Introduce you to simulation-based Bayesian model-fitting using Markov-chain Monte Carlo (MCMC) methods, in the freeware packages WinBUGS and R, when closed-form solutions are not possible.
Course 2 will:
*********************
(1) Introduce Bayesian hierarchical modelling via meta-analysis, the study of how information can be combined across experiments to provide a better summary than those obtained by examining one experiment at a time;
(2) Discuss the critical role played by the choice of prior distributions in Bayesian hierarchical models;
(3) Illustrate the use of latent variables (random effects) as an approach to describing unexplained heterogeneity; and
(4) Explore two in-depth case studies involving random-effects Poisson regression and mixed-effects logistic regression.
Course 3 will:
*********************
(1) Provide an overview of the process of Bayesian model specification;
(2) Introduce five basic principles -- the Calibration Principle, the Modelling-As-Decision Principle, the Prediction Principle, the Inference-Versus-Decision Principle and Cromwell's Rule (Parts 1 and 2) -- and show you how they inform good Bayesian model building and model criticism; and
(3) Introduce you to Calibration Cross-Validation, Bayes factors, BIC, DIC and log scoring (in WinBUGS and R) as methods for finding good Bayesian models.
Course 4 will:
*********************
(1) Introduce you to the discipline of data science, which (broadly speaking) aims to help people find meaningful and verifiable facts and relationships in data sets of enormous size;
(2) Provide a brief introduction to Bayesian non-parametric methods, (a) which offer a flexible approach to model specification that may permit you to avoid false modelling assumptions and (b) which have recently been subject to research that permits them to be used on very large data sets;
(3) Illustrate a variety of best-practice data science methods in a series of real-world case studies ranging in size up to experiments with tens of millions of observations.
All four courses will be based on a series of practical real-world case studies.
*******************************************************************************************************************************************************.
Who should attend?
******************
Statisticians, biostatisticians, epidemiologists, data analysts, data-miners, machine-learning specialists and data scientists who wish to broaden and deepen:
(a) their understanding of Bayesian methods and
(b) their toolkits for using Bayesian models to find meaningful patterns, arrive at statistically sound inferences and predictions, and make better decisions.
Some graduate coursework in statistics (or an allied field such as biostatistics, epidemiology or machine learning) will provide sufficient mathematical background for participants. To get the most out of the course, participants should be comfortable with hearing the course presenter discuss:
(a) differentiation and integration of functions of several variables and
(b) discrete and continuous probability distributions (joint, marginal, and conditional) for several variables at a time.
However, all necessary concepts will be approached in a sufficiently intuitive manner that rustiness on these topics will not prevent understanding of the key ideas.
The first course will assume no previous exposure to Bayesian ideas or methods. Participants interested in entering the second, third or fourth course should ideally have had exposure to the ideas on preceding courses.
*******************************************************************************************************************************************************.
How you will benefit
*******************
You will:
(1) Gain a deeper understanding of maximum-likelihood-based methods and when they can be expected to behave in a sub-optimal manner;
(2) Broaden and deepen your facility in the fitting and interpretation of Bayesian models to solve important problems in science, public policy and business; and
(3) Learn how to write your own programs in WinBUGS and R to fit Bayesian models in your own work.
*******************************************************************************************************************************************************.
Content of courses
***************
Course 1: Bayesian Modelling, Inference, Prediction and Decision-Making
************************************************************
* Background and basics: strengths and weaknesses of the classical, frequentist and Bayesian probability paradigms
* Sequential learning via Bayes' Theorem
* Coherence as a form of internal calibration
* Bayesian decision theory via maximization of expected utility
* Review of frequentist modelling and maximum-likelihood inference
* Exchangeability as a Bayesian concept parallel to frequentist independence
* Prior, posterior, and predictive distributions
* Bayesian conjugate analysis of binary outcomes, and comparison with frequentist modelling
* Conjugate analysis of integer-valued outcomes (Poisson modelling)
* Conjugate analysis of continuous outcomes (Gaussian modelling)
* Multivariate unknowns and marginal posterior distributions
* Introduction to simulation-based computation, including rejection sampling and Markov chain Monte Carlo (MCMC) methods
* MCMC implementation strategies.
Course 2: Bayesian Hierarchical Modelling
***********************************
* Bayesian hierarchical modelling
* Hierarchical modelling with latent variables as an approach to mixture modelling
* Bayesian fixed- and random-effects meta-analysis
* Prior distributions in Bayesian hierarchical modelling
* Bayesian fitting of random-effects and mixed models
* Comparison of likelihood-based and Bayesian methods for fitting hierarchical models: circumstances in which likelihood-based fitting can be poorly calibrated.
Course 3: Bayesian Model Specification
*********************************
* The big picture in Bayesian model specification
* The Calibration Principle, the Modelling-As-Decision Principle, the Prediction Principle, the Inference-Versus-Decision Principle, and Cromwell's Rule (Parts 1 and 2)
* Model expansion as a tool for improving Bayesian modelling: embedding a deficient model in a larger class of models of which it's a special case
* Methods for finding good Bayesian models: Calibration Cross Validation, Bayes factors, BIC, DIC and log scores
* A generic Bayesian model-search algorithm
* False positive/false negative trade-offs in comparing {Bayes factors, BIC} and {DIC, log scores} on their ability to correctly discriminate between models.
Course 4: Studies in Bayesian Small-to-Very-Large-Scale Data Science
********************************************************
* The history of data science and ``Big Data" from 1944 to the present;
* A review of the terminology for data set size, from bytes to yottabytes (2**80 bytes);
* (Bayesian non-parametrics) placing prior distributions on curves, such as cumulative distribution functions (CDFs) and regression surfaces, and following through to posterior distributions on those curves, with particular attention to Dirichlet process priors and how they may be fit in a computationally efficient way, even with gigabyte data sets;
* Large-scale A/B testing, which is the data-science term for randomized controlled trials on customers and web sites: how to do this well, and how to do it badly;
* Design and analysis of very-large-scale observational studies (because you can't always randomize); and
* Methods for simultaneous forecasting of tens of millions of related time series.
******************************************************************************************************************************************************.
Guest Presenter
****************
David Draper is a Professor of Statistics in the Department of Applied Mathematics and Statistics at the University of California, Santa Cruz (USA); he has also been a Senior Principal Research Scientist at both eBay Research Labs and at Amazon Research, where he developed new methods for Bayesian analysis of very large data sets in the discipline of data science.
He is a Fellow of the American Association for the Advancement of Science, the American Statistical Association (ASA), the Institute of Mathematical Statistics, and the Royal Statistical Society; from 2001 to 2003 he served as the President-Elect, President, and Past President of the International Society for Bayesian Analysis (ISBA).
He is the author or co-author of about 150 contributions to the methodological and applied statistical literature, including articles in the Journal of the Royal Statistical Society (Series A, B and C), the Journal of the American Statistical Association, the Annals of Applied Statistics, Bayesian Analysis, Statistical Science, the New England Journal of Medicine, and the Journal of the American Medical Association; his 1995 JRSS-B article on assessment and propagation of model uncertainty has been cited almost 1,500 times, and taken together his publications have been cited almost 13,000 times.
His research is in the areas of Bayesian inference and prediction, model uncertainty and empirical model-building, hierarchical Modelling, Markov Chain Monte Carlo methods, Bayesian nonparametric methods and data science, with applications mainly in medicine, health policy, education, environmental risk assessment and eCommerce.
His short courses have received Excellence in Continuing Education Awards from the American Statistical Association on two occasions, corresponding to days 1 and 2 of this week of courses (21-25 November 2016). He has won or been nominated for major teaching awards everywhere he has taught (the University of Chicago; the RAND Graduate School of Public Policy Studies; the University of California, Los Angeles; the University of Bath (UK); and the University of California, Santa Cruz).
He has a particular interest in the exposition of complex statistical methods and ideas in the context of real-world applications.
Course Materials
****************
Please note that there will not be any printed notes for this course. Sample material from the 2015 courses may be viewed here:
Course 1: Bayesian Modelling, Inference, Prediction and Decision-Making: https://users.soe.ucsc.edu/~draper/Reading-2015-Days-1-and-2.html
Course 2: Bayesian Hierarchical Modelling - https://users.soe.ucsc.edu/~draper/Reading-2015-Day-3.html
Course 3: Bayesian Model Specification - https://users.soe.ucsc.edu/~draper/Reading-2015-Day-4.html
Course 4: Case Studies in Bayesian Data Science in an Era of 'Big Data' - https://users.soe.ucsc.edu/~draper/Reading-2015-Day-5.html
Location
*********
The Statistical Services Centre at Whiteknights campus, University of Reading is in a prime location in the South-East of England and has excellent transport links. The University is close to the M4 motorway allowing easy access by car. Reading's railway station has high speed links to and from London Paddington, as well as regular services to and from other cities around the UK. There are direct services to and from both London Heathrow and London Gatwick Airports. For further details view: http://www.reading.ac.uk/ssc/contact.php.
Kind regards,
James Gallagher
Director, Statistical Services Centre | University of Reading | PO Box 240 | RG6 6FN | United Kingdom
+44 (0)118 378 6730 | www.reading.ac.uk/ssc | [log in to unmask]
-------------------------------------------------------------------
This list is for discussion of modelling issues and the BUGS software.
For help with crashes and error messages, first mail [log in to unmask]
To mail the BUGS list, mail to [log in to unmask]
Before mailing, please check the archive at www.jiscmail.ac.uk/lists/bugs.html
Please do not mail attachments to the list.
To leave the BUGS list, send LEAVE BUGS to [log in to unmask]
If this fails, mail [log in to unmask], NOT the whole list
|