Short course on Flexible Bayesian Methods for Diagnosis and ROC Curve Estimation

Short course on Flexible Bayesian Methods for Diagnosis and ROC Curve Estimation

El curso “Short course on Flexible Bayesian Methods for Diagnosis and ROC Curve Estimation” tendrá lugar en Lisboa del 4 al 6 de junio, impartido por Vanda InácioMiguel de Carvalho (Pontificia Universidad Católica de Chile).

 Program

The course will have a duration of two and a half days and will involve both lectures and practical sessions. The outline is as follows.

1. Introduction and motivation.
1.1. Why are diagnostic tests important?
1.2. Why Bayesian methods?
1.3. Why Bayesian nonparametric methods?

2. Bayesian nonparametric modeling of diagnostic testing data.
2.1. Bayesian principles and methods.
2.2. Polya trees, mixtures of Polya trees and Polya trees mixtures.
2.3. Finite mixtures and Dirichlet process mixtures.
2.4. Hyperpriors.
2.5. Markov chain Monte Carlo approximations to the posterior.

3. Binary and continuous diagnostic tests modeling.
3.1. Techniques for binary tests.
3.2. Diagnosis and ROC estimation: gold standard and no gold standard cases.
3.3. Priors and Identiability.
3.4. Comparison of methods by simulation and data applications.

4. Regression modeling.
4.1. Motivation for inclusion of covariates.
4.2. Dependent mixtures of Polya trees and dependent Dirichlet process mixtures.
4.3. Covariate dependent diagnosis and ROC curve: gold standard and no gold standard cases.
4.4. Priors and identiability.
4.5. Comparison of methods by simulation and data applications.

5. Miscellaneous topics.
5.1. Multivariate diagnosis data.
5.2. Prevalence estimation.
5.3. Sample size determination.

Software: Participants are expected to bring their own laptops with recent versions of R and WinBUGS installed.

Más información: http://ceaul-course-fbm.weebly.com/

in