Hidden Markov Models in Stan

Follow-up

You can find the slides of the workshop here.

Type: WorkshopSpeaker: Vianey Leos BarajasDate & time: Tue May 13 2025, 09:00-12:00 (CEST)Place: Wretlindsalen, Tomtebodavägen 18AURL: https://github.com/vianeylb/LearnBayes

This is an in-person event only. No zoom link will be provided.

Are you working with longitudinal health data, infectious disease surveillance, or patient pathways through complex clinical states? Do your data have gaps, noise, or latent processes that standard methods struggle to capture?

Join us for a hands-on workshop on Bayesian Hidden Markov Models (HMMs) in Stan, tailored specifically for epidemiologists and public health researchers.

Vianey Leos Barajas is an Assistant Professor in Statistical Sciences and affiliated with the School of the Environment at the University of Toronto. Her work focuses topics related to Bayesian statistics, time series, Markov-switching and spatio-temporal models.

Model disease processes more realistically. HMMs allow you to explicitly model unobserved (latent) states — such as infection status, disease stages, or care-seeking behaviors — while accounting for imperfect observations.

No prior Stan experience required. This workshop is designed to be accessible. Whether you’re new to Bayes or Stan, we’ll walk through the basics of HMMs, introduce core Stan concepts, and provide practical examples using real-world epidemiological data.

Hands-on coding and interpretation. You’ll learn not only how to fit HMMs in Stan, but also how to interpret the outputs and assess model fit and identifiability.

Discover new research opportunities. Understanding and applying HMMs can open doors to more rigorous, publishable, and policy-relevant analyses. The skills you’ll gain are highly transferable: you can apply HMMs in a wide range of public health questions involving latent dynamics, from infectious diseases to chronic care modeling.