*** Please note that the hall is CK112.
HIIT seminar, Friday Sep 25, 10:15 a.m. (coffee from 10), Exactum CK112
Dr. Daniel Schmidt
Centre for Molecular, Environmental, Genetic and Analytic Epidemiology The University of Melbourne
The Minimum Message Length Principle for Inductive Inference
Abstract:
The Minimum Message Length (MML) principle is an information theoretic framework for statistical inference proposed and developed by Prof. Chris Wallace and collaborators over the last 50 years.
The MML principle is a Bayesian approach that frames the problem of statistical inference as one of finding a model that leads to the briefest two-part compression of the data, in light of the chosen prior beliefs. While conceptually similar to the related Minimum Description Length (MDL) principle, it differs on a range of philosophical and technical points and for many problems can lead to different answers.
This talk will give an introduction to the MML principle, discuss some of its properties and present several problems to which it has been successful applied.
Bio:
Dr. Daniel Schmidt is a post-doctoral researcher working as a statistician at the Centre for Molecular, Environmental, Genetic and Analytic Epidemiology at The University of Melbourne, Melbourne, Australia. Previously, he completed his PhD in Computer Science at Monash Universityand worked as a post-doctoral researcher at the Faculty of Information Technology, Monash University. His research interests are primarily information theory and its application to statistical inference, with a current emphasis on statistical genetics problems.
Dr. Schmidt is visiting HIIT from 21/9 to 25/9; in case you wish to talk to him, please contact Dr. Teemu Roos, or Daniel directly at dschmidt@unimelb.edu.au.
Last updated on 18 Sep 2009 by Visa Noronen - Page created on 25 Sep 2009 by Visa Noronen