Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach

Estimating the true extent of the outbreak was one of the major challenges in combating COVID-19 outbreak early on. Our inability in doing so, allowed unreported/undetected in- fections to drive up disease spread in numerous regions in the US and worldwide. Accurately identifying the true magnitude of infections still remains a major challenge, despite the use of surveillance-based methods such as serological studies, due to their costs and biases. In this paper, we propose an information theoretic approach to accurately estimate the unreported infections. Our approach, built on top of an existing ordinary differential equations based epi- demiological model, aims to deduce an optimal parameterization of the epidemiological model and the true extent of the outbreak which "best describes" the observed reported infections. Our experiments show that the parameterization learned by our framework leads to a better estimation of unreported infections as well as more accurate forecasts of the reported infec- tions compared to the baseline parameterization. We also demonstrate that our framework can be leveraged to simulate what-if scenarios with non-pharmaceutical interventions. Our results also support earlier findings that a large majority of COVID-19 infections were unreported and non-pharmaceutical interventions indeed helped in mitigating the COVID-19 outbreak.

Download Full-text

On model selection in cosmology

SciPost Physics Lecture Notes ◽

10.21468/scipostphyslectnotes.9 ◽

2019 ◽

Cited By ~ 4

Author(s):

Martin Kerscher ◽

Jochen Weller

Keyword(s):

Model Selection ◽

Bayesian Model ◽

Goodness Of Fit ◽

Ratio Test ◽

Theoretic Approach ◽

Information Theoretic ◽

History Of ◽

The Common ◽

Information Theoretic Approach ◽

The Universe

We review some of the common methods for model selection: the goodness of fit, the likelihood ratio test, Bayesian model selection using Bayes factors, and the classical as well as the Bayesian information theoretic approaches. We illustrate these different approaches by comparing models for the expansion history of the Universe. In the discussion we highlight the premises and objectives entering these different approaches to model selection and finally recommend the information theoretic approach.

Download Full-text

An Information Theoretic Approach to Model Selection: A Tutorial with Monte Carlo Confirmation

Perspectives on Behavior Science ◽

10.1007/s40614-019-00206-1 ◽

2019 ◽

Vol 42 (3) ◽

pp. 583-616 ◽

Cited By ~ 1

Author(s):

M. Christopher Newland

Keyword(s):

Monte Carlo ◽

Model Selection ◽

Theoretic Approach ◽

Information Theoretic ◽

Information Theoretic Approach

Download Full-text

On the Undecidability among Kinetic Models: From Model Selection to Model Averaging

Journal of Cerebral Blood Flow & Metabolism ◽

10.1097/01.wcb.0000050065.57184.bb ◽

2003 ◽

Vol 23 (4) ◽

pp. 490-498 ◽

Cited By ~ 60

Author(s):

Federico E. Turkheimer ◽

Rainer Hinz ◽

Vincent J. Cunningham

Keyword(s):

Model Selection ◽

Model Averaging ◽

Simulated Data ◽

Data Sets ◽

Theoretic Approach ◽

Data Set ◽

Information Theoretic ◽

Physiologic Parameters ◽

Model Set ◽

Information Theoretic Approach

This article deals with the problem of model selection for the mathematical description of tracer kinetics in nuclear medicine. It stems from the consideration of some specific data sets where different models have similar performances. In these situations, it is shown that considerate averaging of a parameter's estimates over the entire model set is better than obtaining the estimates from one model only. Furthermore, it is also shown that the procedure of averaging over a small number of “good” models reduces the “generalization error,” the error introduced when the model selected over a particular data set is applied to different conditions, such as subject populations with altered physiologic parameters, modified acquisition protocols, and different signal-to-noise ratios. The method of averaging over the entire model set uses Akaike coefficients as measures of an individual model's likelihood. To facilitate the understanding of these statistical tools, the authors provide an introduction to model selection criteria and a short technical treatment of Akaike's information–theoretic approach. The new method is illustrated and epitomized by a case example on the modeling of [11C]flumazenil kinetics in the brain, containing both real and simulated data.

Download Full-text