Pre Workshop Friday (9/26) from 1 pm to 2 pm Registration from 2pm to 5 pm - Pre workshop This part of the of workshop will briefly cover basic concepts such as null hypothesis testing, principles of bayesian theory, and maximum likelihood estimation, likelihood function, log-likelihood function, deviance, profile likelihood intervals. Saturday and Sunday (9-27-28) - Workshop from 8 am -12pm and from 1 pm to 5pm Dr. David Anderson will be covering the following topics during Friday and Sunday sections. Science Strategies Based on Kullback-Leibler Information Introduction Philosophy of Science Importance of Models in Science The Concept of Information Overview of the Information-Theoretic Methods Kullback-Leibler Information Information Loss Measuring Information and the Loss Scientific Evidence Chamberlin’s Science Strategy Multiple Working Hypotheses Mapping of Hypotheses and Models Two Science Examples Transmission of TB in Feral Ferrets Bill Lengths in Darwin’s Finches Fundamental Starting Points The Value of the Maximized Log-likelihood Function The Residual Sum of Squares in ‘Regression’ Information Theory Meets Statistical Theory A Glimpse into the Derivation of AIC Expected Kullback-Leibler Information A Small Sample (non-asymptotic) AIC Deltas – Putting Evidence on the Scale of Information Examples of Application Flather’s Landscape Data Cement Data Extending the Theory Likelihood of a Model, Given the Data Model Probabilities (Akaike Weights) Evidence Ratios DURSBAN in a Simulated Ecosystem The Principle of Parsimony Model Bias Model Uncertainty The Trade-off Review of Null Hypothesis Testing Problems Going Back for Nearly a Century Comparison with Information-Theoretic Approaches Multimodel Inference Forms of Model Averaging Unconditional Variances Relative Importance of Variables Second Order Issues Cross validation Overdispersion and QAICc A Likelihood Equivalent of R-sq Model Based Inference from Strict Experiments An AIC for Multivariate Data Summary of Material The theory and application presented will be taken largely from the book, Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel Inference: a practical information-theoretic approach. 2nd Ed. Springer- Verlag, New York, NY. 488pp. David R. Anderson February 9, 2006 |
Support from |
Course Syllabus |