Skip to content

You are here: Home / XV Summer School 2022 / Courses / Introducción a los modelos dinámicos lineales

Introducción a los modelos dinámicos lineales

Date:

June 27 to July 1. AFTERNOON: 15 to 18h

Classroom:

Not defined yet

Instructor

María Eugenia Castellanos Nueda

Currently, I am a Full Professor of Statistics and Operation Research at Rey Juan Carlos University (Madrid), since 2020. Previously, I have been an Assistant professor in Statistics at Miguel Hernandez University (1998-2003). I have also been visiting professor at the Department of Statistics at Carnegie Mellon University, USA (2003); Department of Statistics in Federal University of Rio de Janeiro, Brasil (2007), and several times at the Department of Mathematics at the University of Cagliari, Italy (from 2009 to 2019).

My research mainly focuses on Bayesian statistics from an objective point of view: beginning with the subject of my thesis about measures to quantify the fitting of hierarchical Bayesian models (GOF of Bayesian models), followed by the developing of objective or default priors for modelling extreme values, and continuing with the construction of objective priors for model selection in problems with censored data. All this theoretical development of methodology has been applied also to practical applications with data in real situations. Along with my research career, I have collaborated with experts in different fields: medicine, botanic, anthropology, robotics and informatics. From these collaborations, I managed to publish several articles in journals of different areas, jointly with some multidisciplinary projects, company contracts and some specialized courses. 

I have participated in more than 25 research projects funded by public institutions, being PI in two of them, the last supported by Ministerio de Ciencia Innovación y Universidades. I have also participated in 14 research contracts with companies supervising 7 of them. 

I have been personally invited to give talks in several statistical meetings such as the XVII Spanish Biometric Conference, the International Workshop on Objective Bayes Methodology or at Meetings of the Italian Statistics Society. I have also contributed to scientific conferences in the form of oral discussions, oral presentations and posters.

Since January 2019 I am Associate Editor of Bayesian Analysis (ranked 14/125 Statistics and Probability in JCR 2020). I have also been a referee for journals like JRSS-B, JASA, JSPI, CSDA, Statistics in Medicine, Test, BA, etc. I also reviewed project proposals in the Ministerio de Ciencias, Innovación y Universidades (Spain) and participated in the organization of several meetings like the SEIO 2012 in Madrid (>400) or O-Bayes 2015 in Valencia (>100 participants). I am the chair of the scientific committee of the XVIII Spanish Biometric conference that will take on May 2022.

I am a member of BIOSTATNET, Red Nacional de Bioestadística, since its born in 2010, I am also a member of several societies like ISBA, Biometría and SEIO. I am a member of the research group VaBar (http://vabar.github.io), a top research group in Spain specialising in Bayesian inference. 

Language

Spanish

Description

The course is an introduction to Dynamic Linear Models, beginning with an introduction to Bayesian inference and its notations, and motivating the advantages of using this type of inference to model data over time. All the concepts and models will be introduced with examples in R and using libraries developed to adjust DLMs.

Course goals

The students will know concepts such as Bayesian inference, posterior distribution, dynamic linear model, state-space model. Students will be able to use R functions and libraries to make inferences and predictions with DLMs.

Course contents

  1. Introduction to Bayesian Inference
  2. Definition of a DLM. Examples. 
  3. Simulation of DLMs with R.
  4. Basic DLMs.
  5. Interest problems in DLMs: Filtering, prediction, prediction to k steps, smoothing.
  6. Model specification.
  7. Estimation of unknown parameters. 
  8. Basic Particle filter.

Prerequisites

Statistics and probabilistic concepts.

Targeted at

Students interested in statistics that want to know state-space models

Evaluation

Participation during the class, questionaries, practical questions and data analysis.

Software requirements

R