Welcome

The Workshop on Probabilistic and Statistical Methods is a meeting organized by the Join Graduate Program in Statistics UFSCar/USP (PIPGEs) with the aim of discussing new developments in statistics, probability and their applications.

Main

Activities include invited speaker sessions, short talks, a poster session and a short course devoted to graduate students. The topics of this new edition include probability and stochastic processes, statistical inference, regression models, survival analysis and related topics.

Speakers

Bruno Santos - UFBA
Christian Galarza - UNICAMP
Daniele Granzotto - UEM
Edward George - University of Pennsylvania
Eveliny Barroso da Silva - UFMT
Fernando Quintana - Pontificia U. Católica de Chile
Jean-Yves Dauxois - Université de Toulouse
Jeremias Leão, UFAM
Marcelo Bourguignon Pereira - UFRN
Marina Paez - UFRJ
Paulo Justiniano Ribeiro Junior - UFPR (course)
Nancy Garcia - UNICAMP
Silvia Lopes de Paula Ferrari - USP

Special Sessions

Session: Latent variable modeling (Chair: M. Curi)

Caio Lucidius N. Azevedo - UNICAMP
Hedibert Freitas Lopes - Insper
Jorge Bazán - USP

Session: Probability (Chair: P. Rodriguez)

Carolina Bueno - USP
Cristian Coletti - UFABC
Mary Luz Rodiño - Universidad de Antioquia
Miguel Abadi - USP

Session: Survival Analysis (Chair: V. Tomazella)

Francisco Louzada Neto - USP
Manoel dos Santos Neto - UFSCar and UFCG
Vinicius Calsavara - A. C. Camargo Cancer Center

Committees

Organizing Committee

Daiane Zuanetti - UFSCar
Rafael Stern - UFSCar
Ricardo Ehlers - USP (Chair)
Pablo Rodriguez - USP
Vera Tomazella - UFSCar (Chair)

Scientific Committee
Adriano Polpo - UFSCar
Carlos Alberto de Bragança Pereira - USP
Enrico Colosimo - UFMG
Francisco Louzada Neto - USP
Josemar Rodrigues - USP

Program

See below the titles and abstracts of the conferences and talks.

Conferences

On Bayesian quantile regression

Bruno Santos - UFBA

×

On Bayesian quantile regression

Bruno Santos - UFBA

In this work we discuss the progress of Bayesian quantile regression models since their first proposal and we discuss the importance of all parameters involved in the inference process. Using a representation of the asymmetric Laplace distribution as a mixture of a normal and an exponential distribution, we discuss the relevance of the presence of a scale parameter to control for the variance in the model. Besides that we consider the posterior distribution of the latent variable present in the mixture representation to showcase outlying observations given the Bayesian quantile regression fits, where we compare the posterior distribution for each latent variable with the others. We illustrate these results with simulation studies and also data about Gini indexes in Brazilian states from years with census information.

Mortality rate estimation and standardization for public

Edward George - University of Pennsylvania

× Mortality rate estimation and standardization for public (Edward George - University of Pennsylvania

Bayesian models are increasingly fit to large administrative data sets and then used to make individualized recommendations. In particular, Medicare's Hospital Compare webpage provides information to patients about specific hospital mortality rates for a heart attack or Acute Myocardial Infarction (AMI). Hospital Compare's current recommendations are based on a random-effects logit model with a random hospital indicator and patient risk factors. Except for the largest hospitals, these individual recommendations or predictions are not checkable against data, because data from smaller hospitals are too limited to provide a meaningful check. Before individualized Bayesian recommendations, people derived general advice from empirical studies of many hospitals; e.g., prefer hospitals of type 1 to type 2 because the risk is lower at type 1 hospitals. Here we calibrate these Bayesian recommendation systems by checking, out of sample, whether their predictions aggregate to give correct general advice derived from another sample. This process of calibrating individualized predictions against general empirical advice leads to substantial revisions in the Hospital Compare model for AMI mortality. In order to make appropriately calibrated predictions, our revised models incorporate information about hospital volume, nursing staff, medical residents, and the hospital's ability to perform cardiovascular procedures. For the ultimate purpose of comparisons, hospital mortality rates must be standardized to adjust for patient mix variation across hospitals. We find that indirect standardization, as currently used by Hospital Compare, fails to adequately control for differences in patient risk factors and systematically underestimates mortality rates at the low volume hospitals. To provide good control and correctly calibrated rates, we propose direct standardization instead. This is joint research with Veronika Rockova, Paul Rosenbaum, Ville Satopaa and Jeffrey Silber.

Repulsion in Bayesian mixture models: how and why

Fernando Quintana - Pontificia U. Católica de Chile

×

Repulsion in Bayesian mixture models: how and why

Fernando Quintana - Pontificia U. Católica de Chile

Bayesian mixture models have become very popular for applications such as density estimation and clustering. A typical assumption in this context is that component-specific parameters are modeled as independent quantities. An undesired consequence of this assumption in the context of clustering is the presence of very small or singleton clusters, which are very hard or impossible to interpret. This talk presents some approaches that have been recently proposed for solving this problem. The basic idea is to introduce the notion of repulsion of component-specific location parameters, that is, a probability model that encourages separation of these quantities. Specific constructions and applications to density estimation and clustering will be discussed.

Statistical inference in models of imperfect maintenance with geometric or arithmetic reduction of intensity

Jean-Yves Dauxois - Université de Toulouse

×

Statistical inference in models of imperfect maintenance with geometric or arithmetic reduction of intensity

Jean-Yves Dauxois - Université de Toulouse

In this talk I will introduce and study two new models of Imperfect Maintenance in Reliability: a model of Geometric Reduction of Intensity and another of Arithmetic Reduction of Intensity on the inter-arrival times of failures on a system subject to recurrent failures. Based on the observation of the recurrent failures of a single repairable system and assuming that a perfect repair is operated after N failures, we introduce estimators of the parameters (euclidean and functional) in this two semiparametric models and we prove their asymptotic normality. Then a simulation study is carried out to learn the behavior of these estimators on samples of small or moderate size. We will end with applications on a real dataset.

An INAR(1) process for modeling count time series with equidispersion, underdispersion and overdispersion

Marcelo Bourguignon Pereira - UFRN

×

An INAR(1) process for modeling count time series with equidispersion, underdispersion and overdispersion

Marcelo Bourguignon Pereira - UFRN

We present a novel first-order non-negative integer-valued autoregressive model for stationary count data processes with Bernoulli-geometric marginals based on a new type of generalized thinning operator. It can be used for modeling time series of counts with equidispersion, underdispersion and overdispersion. The main properties of the model are derived, such as probability generating function, moments, transition probabilities and zero probability. The maximum likelihood method is used for estimating the model parameters. The proposed model is fitted to time series of counts of iceberg orders and of cases of family violence illustrating its capabilities in challenging cases of overdispersed and equidispersed count data. Joint work with: Christian H. Weiß - Department of Mathematics and Statistics, Helmut Schmidt University, Hamburg, Germany.

Hierarchical stochastic block model for community detection in multiplex networks

Marina Paez - UFRJ

×

Hierarchical stochastic block model for community detection in multiplex networks

Marina Paez - UFRJ

Multiplex networks have become increasingly more prevalent in many fields, and emerged as a very powerful tool for modeling the complexity of real networks. There is a critical need for developing statistical models for inference in multiplex networks that can take into account potential dependency across different layers. There is in particular a demand for models for community detection. We fill this gap by proposing a novel and efficient Bayesian model for community detection in multiplex networks that take into account the dependency within and across different layers. A random partition prior is imposed for partitions across different layers of the multiplex network, under which a stochastic block model (SBM) is assumed. We also assume that the structure of the partitions is somewhat similar by imposing a hierarchical stochastic block model (HSBM) to the multiplex network. One of the key features of our model is that it allows the communities at different layers of the network to vary, which differs from many of existing methods for modeling multiplex networks, which assume that the communities are the same or fixed for all the layers. Efficient MCMC algorithms were developed for sampling the posterior of `communities', or the partition structure, as well as the link probabilities between nodes or communities. The developed algorithms were applied to extensive simulation studies and data examples which demonstrated the good performance of the models and algorithms.

Modeling textile images with hidden Gibbs random fields

Nancy Garcia - UNICAMP

×

Modeling textile images with hidden Gibbs random fields

Nancy Garcia - UNICAMP

When a new textile dyeing technology is developed, evaluating the quality of these techniques involves measuring the resulting color homogeneity using digital images. The presence of a texture caused by the fabric creates a sophisticated dependence structure in pixels coloring that is not accommodated by the available probabilistic models. Due to several factors, the random field that generates the texture can be seen as a mixture of colors and the mixture is given by a hidden Gibbs process with complex interactions. Joint work with Victor Freguglia Souza

Box-Cox t random intercept model for estimating usual nutrient intake distributions

Silvia Lopes de Paula Ferrari - USP


×

Box-Cox t random intercept model for estimating usual nutrient intake distributions

Silvia Lopes de Paula Ferrari - USP

The issue of estimating usual nutrient intake distributions and prevalence of inadequate nutrient intakes is of interest in nutrition studies. Box-Cox transformations coupled with the normal distribution are usually employed for modeling nutrient intake data. When the data present highly asymmetric distribution or include outliers, this approach may lead to implausible estimates. Additionally, it does not allow interpretation of the parameters in terms of characteristics of the original data and requires back transformation of the transformed data to the original scale. We propose an alternative approach for estimating usual nutrient intake distributions and prevalence of inadequate nutrient intakes through a Box-Cox t model with random intercept. The proposed model is flexible enough for modeling highly asymmetric data even when outliers are present. Unlike the usual approach, the proposed model does not require a transformation of the data. A simulation study suggests that the Box-Cox t model with random intercept estimates the usual intake distribution satisfactorily, and that it should be preferable to the usual approach particularly in cases of highly asymmetric heavy-tailed data. In applications to data sets on intake of 19 micronutrients, the Box-Cox t models provided better fit than its competitors in most of the cases. Joint work with Giovana Fumes and José Eduardo Corrente

Mini-Conferences

On moments of truncated multivariate Student-t distribution: a recurrence approach

Christian Galarza - UNICAMP

×

On moments of truncated multivariate Student-t distribution: a recurrence approach

Christian Galarza - UNICAMP

Recurrence relations for integrals that involve the density of multivariate Student-t distributions are developed. These recursions allow fast computation of the moments of folded and truncated multivariate normal and Student-t distributions. Besides being numerically efficient, the proposed recursions also allow us to obtain explicit expressions of low order moments of folded and truncated multivariate Student-t distributions. The newly methods are implemented in the new R package MoMtt. Joint work with: Victor Hugo Lachos Davila - UNICAMP, Tsung-I Lin - National Chung Hsing University (Taiwan) and Wan-Lun Wang - Feng Chia University (Taiwan).

Gompertz-log-logistic distributions and minimum quadratic distance estimation

Daniele Granzotto - UEM

×

Gompertz-log-logistic distributions and minimum quadratic distance estimation

Daniele Granzotto - UEM

The objective of this study is twofold: first to introduce two new parametric families of distributions with support on the positive real line that contain both the Gompertz and the log-logistic distributions as particular cases. The new families are flexible as they contain increasing, decreasing and bathtub or inverted bathtub hazard rate functions. The construction of the new families is based on the cumulative hazard function and this leads to the second contribution of the research, a new method of parameter estimation based on minimizing a quadratic distance from the cumulative hazard of the model to the Nelson- Aalen estimator. We illustrate the use of the new estimation method on the proposed families and performed a simulation study to illustrate the finite sample properties of the derived estimators to show some of the robustness properties of the minimum quadratic distance estimators as compared to maximum likelihood. Joint work with: Karim Anaya-Izquierdo - University of Bath and Francisco Louzada of USP.

Beta regression model with multiplicative measurement errors: an application to financial data*

Eveliny Barroso da Silva - UFMT

×

Beta regression model with multiplicative measurement errors: an application to financial data

(*Canceled)

Eveliny Barroso da Silva - UFMT

The construction of a beta regression model with multiplicative measurement error is motivated by problems in the financial area. Financial institutions need to determine the proportion of spending of the credit limit to be offered at the special check for future clients. The knowledge of this proportion allows the institution to allocate an amount of capital to cover a possible default risk. An important factor in determining the proportion of spending is the customer's income. Since we are dealing with new clients, information about their real income is not usually available and the institution obtains their presumed income on the market, that is, an income measured with error. In this work, we present the beta regression model with log-normal multiplicative error. Some estimation methods are studied and compared. These methods have as principle the estimation by maximum likelihood and pseudo likelihood. A simulation study was carried out to illustrate the results of the estimates for each method. In addition, we analyze a real data and perform diagnostic analysis. This is a joint work with Carlos Alberto Ribeiro Diniz (UFSCar, Brazil).

Birnbaum-Saunders frailty regression models: diagnostics and application to medical data

Jeremias Leão, UFAM


×

Birnbaum-Saunders frailty regression models: diagnostics and application to medical data

Jeremias Leão, UFAM

In survival models, some covariates affecting the lifetime could not be observed or measured. These covariates may correspond to environmental or genetic factors and be considered as a random effect related to a frailty of the individuals explaining their survival times. We propose a methodology based on a Birnbaum-Saunders frailty regression model, which can be applied to censored or uncensored data. Maximum-likelihood methods are used to estimate the model parameters and to derive local influence techniques. Diagnostic tools are important in regression to detect anomalies, as departures from error assumptions and presence of outliers and influential cases. Normal curvatures for local influence under different perturbations are computed and two types of residuals are introduced. Two examples with uncensored and censored real-world data illustrate the proposed methodology. Comparison with classical frailty models is carried out in these examples, which shows the superiority of the proposed model.

Short Course

Métodos computacionais para inferência estatística

Paulo Justiniano Ribeiro Junior - UFPR


×

Métodos computacionais para inferência estatística

Paulo Justiniano Ribeiro Junior - UFPR

Métodos numéricos e/ou computacionais são parte essencial do ferramental para tratamento de problemas de inferência estatística. Ambientes de prototipação e linguagens como o R (www.r-project.org) favorecem a implementação de métodos visando flexibilidade na especificação de modelos estatísticos. Este curso visa revisar a discutir diversos métodos aplicáveis em modelagem estatística. Serão abordados procedimentos numéricos como algorítmos para otimização/maximização, solução de sistemas, dentre outros, bem como métodos estocásticos para inferência via métodos computacionalmente intensivos. Exemplos de implementação são fornecidos em linguagem R.

Schedule

See below the poster of the meeting. A full list of titles and abstracts of the oral communications and posters will be updated soon.

Schedule

Click here to download the schedule of the workshop (updated on Jan 18).

Special Session: Latent variable modeling

TBA

Special Session: Probability

TBA

Special Session: Survival Analysis

TBA

Oral Communications

Abstract submissions are received until January 19th 2018

Poster presentations

Abstract submissions are received until January 19th 2018




Participants

List of registered participants. Last update of the list on January 18th, 2018.
Remember that all attendees must pay the appropriate registration fee (see the Information section below for further details).

Adriano Kamimura Suzuki, ICMC-USP
Alana Gabriela Salatim Novais, UFSCar
Alex Rodrigo dos Santos Sousa, UNICAMP
André Alves Ambrósio, ICMC-USP
Anna Caroline, UFEs
Bruno Santos, UFBA
Caio Lucidius N. Azevedo, UNICAMP
Camila Lorencetti Brolo, UFSCar
Carolina Bueno, USP
Christian Galarza, UNICAMP
Claudia Evelyn Escobar Montecino, USP/UFSCar
Clécio da Silva Ferreira, UFJF
Cristian Favio Coletti , UFABC
Daiane de Souza Santos, USP/UFScar
Daiane Zuanetti, UFSCar
Daniele Granzotto, UEM
Davi Keglevich Neiva, UFSCar
Démerson André Polli, UFSCar/ICMC-USP
Diego Carvalho do Nascimento, USP/UFSCar
Diogo Barboza Moreira, UFSCar
Edward George, University of Pennsylvania
Fabiana Arca Cruz Tortorelli, ICMC-USP/UFSCar
Fernanda Rodrigues Vargas, UFRGS
Fernando Quintana, Pontificia U. Católica de Chile
Francisco Antonio Loyola Lavin, USP
Gabriel Marcelino Alves, UFSCar
Gabriela Cintra Raquel, USP/UFSCar
Gesiel Rios Lopes, ICMC-USP

Gianpedro Robertto Mella Brigante, USP
Hedibert Freitas Lopes, Insper
Jean-Yves Dauxois, Université de Toulouse
Jeremias Leão, UFAM
Joao Carlos Poloniato Ferreira, UFSCar/USP
Jorge Bazán, USP
José Fausto de Morais, UFU
Juliana Cobre, ICMC/USP
Juliana Marambaia Maia, ICMC-USP/UFSCar
Juliana Scudilio Rodrigues, UFSCar/ICMC-USP
Karlla Delalibera Chagas, FCT-UNESP
Karoline Eduarda Lima Santos, USP
Lorena Yanet Cáceres Tomaya, UFSCar/USP
Lucas Antonio Barbano, UFSCar
Lucas Eduardo de Moraes, UFSCar
Lucas Leite Cavalaro, DEs-UFSCar/ICMC - USP
Luciana Moura Reinaldo, UFC
Luiz Carlos Medeiros Damasceno, UFSCar/USP
Luiz Gabriel Fernandes Cotrim, UFSCar/ICMC-USP
Maira Fatoretto, ESALQ -USP
Manoel dos Santos Neto, UFSCar/UFCG
Marcelo Bourguignon Pereira, UFRN
Marcos Jardel Henriques, UFSCar-USP
Marina Paez, UFRJ
Mário de Castro, USP
Mary Luz Rodiño, Universidad de Antioquia
Miguel Abadi, USP
Milene Alves Garcia, USP/UFSCar

Milton Miranda Neto, UFSCar/USP
Murilo Cantoni, UFSCar-USP
Nancy Garcia, UNICAMP
Nicholas Wagner Eugenio, IME-USP
Oilson Alberto Gonzatto Junior, UFSCar-USP
Pablo Rodriguez, USP
Paulo Justiniano Ribeiro Junior, UFPR
Rafael Izbicki, UFSCar
Rafael Stern, UFSCar
Renan de Padua, ICMC-USP
Renata Porto Sampaio, Brazil
Renato Gava, UFSCar
Ricardo De Carli Novaes, USP
Ricardo de Jesus Caldas Assis, USP/UFSCar
Ricardo Ehlers, USP
Ricardo Felipe Ferreira, UFSCar/USP
Robinson Nelson dos Santos, Springer Nature
Silvia Lopes de Paula Ferrari, USP
Tainá Santana Caldas, UFRR
Taís Roberta Ribeiro, USP/UFSCar
Tatiane Carvalho Alvarenga, UFLA
Teh Led Red, USP
Thiago Fernando Ferreira Costa, USP
Tiago Bernardes Kerr, UFU
Vanessa Rufino da Silva, USP-ICMC/UFSCar
Vera Tomazella, UFSCar
Vinicius Calsavara, A. C. Camargo Cancer Center




Information

For additional information, please contact us here. We will contact you back as soon as possible.

Important Dates

Deadline abstract submission: Jan 19th
Notification of acceptance: Jan 22th
Deadline for early registration: Jan 26th

Early Fees* (before/on Jan 26th)

Researchers/others: R$ 50,00
Graduate students: R$ 35,00
Undergraduate students: R$ 10,00

Regular Fees** (after Jan 26th)

Researchers/others: R$ 60,00
Graduate students: R$ 45,00
Undergraduate students: R$ 20,00

Where at UFSCar

Registration and main activities will be held at the Anfiteatro Bento Prado Junior. The Special Sessions will be held at the Auditórios 1-3, Biblioteca Comunitária.
See here the campus map for locations.


*Early fees must be paid at Associação Brasileira de Estatística CNPJ 56572456/0001-80: Banco Santander (033), Agência 0658, C/C 13006798-9, and the receipt must be sent to secretaria@redeabe.org.br with the subject [pagamento 6WPSM]. (O pagamento antecipado deve ser realizado por transferência bancária para Associação Brasileira de Estatística CNPJ 56572456/0001-80: Banco Santander (033), Agência 0658, C/C 13006798-9, e o comprovante do depósito deve ser enviado por e-mail para secretaria@redeabe.org.br com o assunto [pagamento 6WPSM])
**Regular fees may be paid in cash at the registration desk on the first day.

Organization


      


Support


                                         

The meeting in numbers

Since the first edition of the Workshop on Probabilistic and Statistical Methods, helded in 2013, we have received the contribution of many colleagues and students.

Conferences and Talks
Oral Communications
Poster Presentations
Participants

Questions?

If you have any question, please do not hesitate in contact us! We are looking forward to seeing you in São Carlos.

Contact Info

E-mail: wpsm@icmc.usp.br

PIPGEs at UFSCar
DEs-Universidade Federal de São Carlos
Rod. Washington Luiz, km 235
CEP: 13565-905 - São Carlos - SP - Brazil

PIPGEs at USP
ICMC - Universidade de São Paulo
Av. Trabalhador São-carlense, 400, Centro
CEP: 13566-590 - São Carlos - SP - Brazil