Ohio State is in the process of revising websites and program materials to accurately reflect compliance with the law. While this work occurs, language referencing protected class status or other activities prohibited by Ohio Senate Bill 1 may still appear in some places. However, all programs and activities are being administered in compliance with federal and state law.

Seminar: Lam Si Tung Ho

Statistics Seminar Series
February 4, 2016
All Day
209 W. Eighteenth Ave. (EA), Room 170

Title

Direct Likelihood-Based Inference for Stochastic Models of Disease Dynamics

Speaker

Lam Si Tung Ho, University of California, Los Angeles

Abstract

Two big data scenarios challenge direct inference for stochastic models: (1) the massive amount of measurements overwhelm traditional methods, and (2) the size of unmeasurable data causes intractability in likelihood calculation.
 
Observational health databases present both massive resources and challenges for studying clinical practices. I will show how to leverage these data by using birth-death processes as a scalable model for treatment trajectories in diabetic patients. Applying this model to 1.6 million diabetic patients from 4 national databases raises questions about the mismatch between treatment guidelines and practice.
 
On the hand, infectious diseases epidemics often have a large number of unmeasurable data. Specifically, it is usually impossible to know the exact infection and removal times for each individual. Therefore, likelihood calculation for prevalence count data under the susceptible-infected-removed (SIR) model has been considered intractable since its formulation almost a century ago. To circumvent this intractability problem, researchers often rely on data augmentation or approximation that may grow computationally expensive or lack accuracy. I will present an efficient and scalable method to directly compute the likelihood function under small stochastic compartmental models such as SIR and apply this method to draw direct inference about the 17th century Great Plague in Eyam and the 2014–15 Ebola outbreak in Guinea.