Datas:
1-3 June 2022
Local:
Faculty of Sciences – University of Lisbon (Faculdade de Ciências da Univesidade de Lisboa),
Campo Grande
1749-016 Lisboa Portugal
T (+351) 217 500 000
EVENT SPEAKER:
Dr. Paciorek is an expert in Bayesian and spatial statistics applied to environmental and public health research. He is an adjunct Professor and statistical computing consultant in the Department of Statistics at the University of California, Berkeley.
In 2003, Dr. Paciorek finished his PhD in Statistics at Carnegie Mellon University. He continued his career with a role as an assistant professor in the Biostatistics Department at the Harvard University School of Public Health.
Since 2009, Dr. Paciorek has been in the Statistics Department at UC Berkeley. In addition to his current positions as adjunct Professor and statistical computing consultant, he is also co-Principal Investigator of the NIMBLE project. He has worked on various aspects of NIMBLE development, including model processing, MCMC and other algorithms. He has also led or co-led a variety of workshops on NIMBLE (hyperlink to https://github.com/nimble-training). His applied statistics work with NIMBLE has included prediction of past vegetation using paleoecological proxy data, hidden Markov modeling of precipitation, and clustering of countries based on health metrics.
More information:
https://www.stat.berkeley.edu/~paciorek/
About:
This workshop is organized by Centro de Estatística e Aplicações (CEAUL) and will give you a better understanding of how to use and customize NIMBLE’s statistical algorithms.
NIMBLE is a system built in R, that allows you to build and analyze statistical models. It is specifically relevant to hierarchical models and computationally-intensive methods and can aid in efficiency by reducing and simplifying your work.
No previous NIMBLE experience is required for this workshop, but a basic understanding of R is advised.
Program:
Day 1 (9h-12h and 14h-17h):
– Introduction to NIMBLE
– Writing models in NIMBLE
– Comparing and customizing MCMC methods
– Strategies for improving MCMC
– Writing your own functions and distributions for NIMBLE models
Day 2 (9h-12h and 14h-17h):
– Introduction to programming algorithms (using nimbleFunctions) in NIMBLE
– Model selection and Bayesian nonparametrics
– Spatial modeling and state-space models
– Sequential Monte Carlo and particle MCMC
Day 3 (half day) (9h-12h):
– Advanced algorithm programming (writing your own MCMC sampler, calling out to R and C++, and more)
– Special topics based on participant interests and discussion of participants’ research projects
For more information, click here
Referência Projeto: Projecto FCT: UIDB/00006/2020