2022

Título: Asymptotic behavior of temporal aggregation in mixed-frequency datasets

Resumo: SHerein, we present an unexplored issue with temporal aggregation; namely, when a model contains frequency-dependent coefficients, such as a distinct long- and short-term coefficient, temporal aggregation leads to inconsistent least squares estimators. Because the subsampled variable’s spectrum is equal to its folded original spectrum, the low-frequency variable may exhibit a mixture of distinct linear relations for a given frequency. We propose a new method to disentangle frequencies superposition based on band spectrum regression, thus avoiding the inconsistency problem. As a result, we can test for the presence of frequency-dependent coefficients. We use stationary and nonstationary linear semiparametric models to demonstrate our findings. Our Monte Carlo simulations show good finite sample size and power properties. Finally, we present an empirical application of our method using quarterly GDP and monthly indicators from the United States.

Título: MANAGERS VERSUS MACHINES: DO ALGORITHMS REPLICATE HUMAN INTUITION IN CREDIT RATINGS?

Resumo: We use machine learning techniques to investigate whether it is possible to replicate the behavior of bank managers who assess the risk of commercial loans made by a large commercial US bank. Even though a typical bank already relies on an algorithmic scorecard process to evaluate risk, bank managers are given significant latitude in adjusting the risk score in order to account for other holistic factors based on their intuition and experience. We show that it is possible to find machine learning algorithms that can replicate the behavior of the bank managers. The input to the algorithms consists of a combination of standard financials and soft information available to bank managers as part of the typical loan review process. We also document the presence of significant heterogeneity in the adjustment process that can be traced to differences across managers and industries. Our results highlight the effectiveness of machine learning based analytic approaches to banking and the potential challenges to high-skill jobs in the financial sector..

Título: Evaluating Markov chain Monte Carlo for phylogenetics using exchangeable distributions

Resumo: Clades are a fundamental object in phylogenetics, being the building blocks of phylogenies. Zhu et al. (2011) derive the joint distribution of clades under two important phylogenetic models, the Yule-Harding-Kingman (YHK) model and the proportional to distinguishable arrangements (PDA) model, which are known as exchangeable phylogenetic distributions (Aldous, 1996). In this talk we build on the results of Zhu et al. (2011) to study the correlation matrix between clade indicators in exchangeable distributions. We then show how to use these theoretical results for simultaneously assessing correctness and performance of phylogenetic MCMC samplers.

Título: CovidLP: um aplicativo para prever a evolução da COVID-19 baseado nos dados

Resumo: Nesta palestra, pretendo apresentar o processo de desenvolvimento de um aplicativo Shiny para previsão de curto e longo prazos para a pandemia da COVID-19. Essas previsões são completamente baseadas nos dados observados diariamente de casos confirmados e mortes, e permitem estimar características relevantes da pandemia, como pico e fim dos casos/mortes e número total de casos/mortes. Toda a inferência é sumarizada em preditores pontuais, acompanhados dos respectivos intervalos de credibilidade para intervalos de curto e longo prazo. Desafios associados ao desenvolvimento de um sistema de larga escala para vários países e estados também serão descritos, além dos produtos resultantes como um pacote do R e o livro “Building a Platform for Data-Driven Pandemic Prediction” publicado em 2021. Esse projeto foi desenvolvido em conjunto com Dani Gamerman, Marcos Prates, Vinicius Mayrink, Guido Moreira e um grupo de alunos e ex-alunos dos Programas de Pós-Graduação em Estatística da UFMG e da UFRJ.