| 20 octubre, 2015 | ||
| 12:00 | a | 12:45 |
| 12:45 | ||
SEMINARIOS DEL DOCTORADO DE ESTADÍSTICA, OPTIMIZACIÓN Y MATEMÁTICA APLICADA Y CONMEMORACIÓN 10º ANIVERSARIO DE LA CREACIÓN DEL INSTITUTO CIO DE LA UMH. CURSO 2015-2016
Título Seminario 1: Revisiting Bayesian hierarchical models, a tool to approach many highly complicated real systems
Ponentes: Prof. Dr. David V. Conesa Guillén (Universitat de València)
Fecha: martes 20 de octubre de 2015 a las 12:00 horas
Lugar: Sala de Seminarios del Instituto Universitario de Investigación CIO, Edificio Torretamarit
Resumen:
In this talk, we will review one of the most important tools of Bayesian analysis, namely the Bayesian hierarchical models. This kind of models are becoming so popular when one has to deal with many real situations such as those appearing in the omics fields, Epidemiology, where it can be natural to have tens or hundreds of thousands of variables to be analysed. Bayesian hierarchical models have been used largely but now they have become kind of necessary in all those practical complex situations because of their ability to deal with them. After reviewing the models, we will show some of the numerical approaches that have been introduced to deal with them (simulation methods such as MCMC, or Laplace approximations, such as INLA), and finally some examples of their applications in various practical settings will be presented.
Breve Bio:
David Conesa Guillén es Profesor Titular del departamento de Estadística e I.O. en la Universitat de València. Ha publicado numerosos artículos sobre modelización Bayesiana, es editor de las dos revistas españolas de estadística: SORT y TEST. Actualmente ostenta el cargo de Presidente de la Sociedad Española de Biometría. Ha impartido cursos sobre esta temática tanto en universidades nacionales como extranjeras.
***
Título Seminario 2: Bayesian spatio-temporal approach in the context of species distribution modelling
Ponentes: Prof. Dr. Maria Grazia Pennino (Universidade Federal do Rio Grande do Norte, Natal (Brasil))
Fecha: martes 20 de octubre de 2015 a las 12:45 horas
Lugar: Sala de Seminarios del Instituto Universitario de Investigación CIO, Edificio Torretamarit
Resumen:
Modelling patterns of occurrence and abundance of the species using local environmental factors has been a growing problem in Ecology in the last few years. This kind of modelling has been extensively used to address several issues, including identifying Essential Fish Habitats (EFHs) in order to classify and manage conservation areas, and predicting the response of species to environmental features.
Different approaches and methodologies have been proposed for modelling the distribution of species. Generalized linear and additive models (GLM and GAM), species envelope models such as BIOCLIM, neural networks and the multivariate adaptive regression splines (MARS) are some of them.
The main issues are that most of these applications are only explanatory and ignore spatial correlations, leading to misleading results.
In this talk, we present how hierarchical models can be used to predict the distribution of species in an accurate way by dealing with the mentioned issues in a more appropriate way.
Breve Bio:
Maria Grazia Pennino, Biologa Marina, formada como investigadora entre el Instituto Oceaongráfico Español y la Universitat de València, actualmente se encuentra realizando estancia post-doctoral de 2 años en la Universidade Federal do Rio Grande do Norte, Natal (Brasil); aunque con anterioridad estuvo contratada dos años en el IRD Institut de Recherche pour le Developpement, Paris, Francia. Es autora de distintas publicaciones sobre la distribución de especies tanto en revistas de Biología Marina como en revistas de Estadística.
Eventos del EOMA