2015-03-20

Dan Nettleton. Iowa State University
Case-Specific Random Forests
Sala 2 Facultad de Matemáticas - 12:00 Hrs.

2015-03-06

Michal Abrahamowicz. Mcgill University
Flexible Modeling of The Cumulative Effects of Time-Varying Covariates in Survival Analyses
Sala 2 Facultad de Matemáticas - 12:00 Hrs.

2015-01-15

Marc G Genton. King Abdullah University of Science and Technology
Visualization and Ranking of Functional Data
Sala 5 Facultad de Matemáticas - 12:00 Hrs.

2015-01-15

Ying Sun. King Abdullah University of Science and Technology
Statistically and Computationally Efficient Estimating Equations for Large Spatial Datasets
Sala 1 Facultad de Matemáticas 12:30 Hrs

2015-01-14

Zhuoqiong He. University of Missouri
Adjusting Nonresponse Bias in Small Area Estimation Via Bayesian Hierarchical Spatial Models
Sala 1 Facultad de Matemáticas - 12:00 hRS.

2015-01-09

Dongchu Sun. University of Missouri
Bayesian Analysis of Multivariate Smoothing Splines
Sala 1 Facultad de Matemáticas - 12:00 Hrs.

2014-12-19

Paul Doukhan. University of Cergy-Pontoise
Using Weak Dependence Tools for Dealing With Extreme Values of Time Series
Sala 1 Facultad de Matemáticas a las 12:00 Hrs.

2014-11-28

Emilio Porcu. Universidad Federico Santa Maria
The History and Mysteries Surrounding $Omega_D$
Sala 1 Facultad de Matemáticas - 12:00 Hrs.

2014-11-14

Ronny O. Vallejos. Universidad Técnica Federico Santa María
Assessing The Similarity Between Images
Sala 1 Facultad de Matemáticas - 12:00 Hrs.

2014-10-03

Mauricio Castro. Universidad de Concepción
Likelihood-Based Inference for Tobit Confirmatory Factor Analysis Using The Multivariate Student-T Distribution
Sala 1 Facultad de Matemáticas

2014-09-12

Luis Gutiérrez. Escuela de Salud Pública, Facultad de Medicina, Universidad de Chile
Statistical Shape Analysis Based on Parameterized Closed Curves
Sala 1 Facultad de Matemáticas a las 12:00 Hrs.

2014-05-30

Emilio Porcu. Universidad Tecnica Federico Santa Maria
Gaussian Fields Evolving Temporally Over Planet Earth
Sala 1 Facultad de Matemáticas - 12:00 Hrs.

2014-05-23

Federico Crudu. Pontificia Universidad Católica de Valparaiso
Jackknife Instrumental Variable Estimation With Heteroskedasticity
Sala 1 - Facultad de Matemáticas - 12:00 Hrs.

2013-11-22

Ricardo Olea. Pontificia Universidad Católica de Chile
Procesos Localmente Estacionarios en Anillos de Crecimiento
Sala 1 - Facultad de Matemáticas - 12:00 Hrs.

2013-10-11

Manuel Galea. Pontificia Universidad Católica de Chile
Diagnósticos de influencia en análisis multivariado bajo distribuciones no-normales
Sala 1 - Facultad de Matemáticas - 12:00 Hrs
Abstract:
Abstract:

El principal objetivo de este trabajo es discutir diagnósticos de influencia en modelos simétricos multivariados. Se discuten algunas técnicas de eliminación de casos (distancias de Cook y de Frechet) y el método de influencia local. Finalmente se ilustra la metodología con datos reales.

2013-09-27

Mauricio Álvarez. Departamento de Ingeniería Eléctrica, Universidad Tecnológica de Pereira, Colombia
Multi-output Gaussian Processes, and applications in dynamical systems
Sala 1 - Facultad de Matemáticas
Abstract:
Abstract:
In this talk, we will review the problem of modeling correlated outputs using Gaussian process priors. Applications of modeling correlated outputs include the joint prediction of pollutant metals in geostatistics, and multitask learning in machine learning. Defining a Gaussian process prior for correlated outputs translates into specifying a suitable covariance function that captures dependencies between the different output variables. Classical models for obtaining such a covariance function include the linear model of coregionalization and process convolutions. We describe a general framework for developing multiple output covariance functions by performing convolutions between smoothing kernels particular to

2013-09-06

Adam Branscum. School of Biological and Population Health Sciences, Oregon State University
New Developments in Bayesian Semiparametric Regression
Sala 1 - Facultad de Matemáticas
Abstract:
Abstract:

A new method is introduced for risk regression with continuous response data that simultaneously provides a goodness of fit test of logistic regression and the opportunity for semiparametric estimation of risks, risk ratios, and odds ratios using Polya trees. Computational methods for empirical and fully Bayesian inference are detailed, and theoretical results establishing the consistency of an empirical Bayes goodness of fit test are examined. Then, dependent Polya trees are used to develop a semiparametric regression model that provides flexibility to accommodate evolving residual distributions. The methodology can be used in a wide range of regression models, such as linear and nonlinear fixed-effects, random-effects, and

2013-07-12

Marta Pérez-Casany. Departamento de Matemática Aplicada Ii, Universidad Politécnica de Catalunya, España
Distribuciones discretas con soporte los enteros no negativos
Sala 1 - Facultad de Matemáticas - 12:00 Hrs.
Abstract:
Abstract:
Los datos correspondientes a contages son requeridos en áreas de investigación muy diversas.
Estos datos tienen unas características propias que se repiten con más o menos frecuencia en los diferentes ámbitos de investigación, y que dependen en gran medida del tipo de población mostreada, así como del mecanismo utilizado para obtener las observaciones de la variable objeto de estudio. El objetivo de la presentación es el de exponer las peculiaridades de la modelización de datos correspondientes a contages, así como el de resaltar los modelos estadísticos más apropiados en cada situación.

2013-06-07

Garritt L. Page. Depto Estadística, Facultad de Matemáticas, Pontificia Universidad Católica de Chile
Clustering heterogeneous functions via shape and subject-specific covariates
Sala 1 - Facultad de Matemáticas - 12:00 Hrs.
Abstract:
Abstract:
We consider multi-subject studies where units of observation are functional curves that are realizations of measurements taken on individuals or experimental units typically over time. Often, in this setting, the observed functional output displays large amounts of between subject heterogeneity in the sense that some individuals produce curves that are fairly smooth while others are (much) more erratic. We argue that this variability in curve shape is a feature that can be exploited to guide decision making or learn about processes under study. In this paper we develop a methodology that takes advantage of this feature when clustering functional curves. Individual curves are flexibly modeled

2013-05-31

ArnošT Komárek. Depto de Probabilidad y Estadística Matemática, Facultad de Matemáticas y Física, Universidad Carolina de Praga, República Checa
Clustering for multivariate continuous and discrete longitudinal data
Sala 1 - Facultad de Matemáticas - 12:00 Hrs.
Abstract:
Abstract:

Multiple outcomes, both continuous and discrete are routinely gathered on subjects in longitudinal studies. We propose a model-based statistical method for clustering (classification) of subjects into a prespecified number of groups with a priori unknown characteristics on basis of repeated measurements of all longitudinal outcomes and show its implementation in the R package mixAK. We start by modeling the evolution of each outcome using the classical generalized linear mixed model (GLMM). Possible dependence between the values of different outcomes is captured by specifying a joint distribution of all random effects involved in the GLMM for each o