On September 19, at 4 pm, the FGV´s School of Applied Mathematics (FGV EMAp) will hold a seminar called “A dynamic Bayesian nonparametric autoregressive model for multiple time series analysis.” The event will take place in Auditorium 537 in FGV’s main building (Praia de Botafogo, 190, Botafogo, Rio de Janeiro).

The seminar will debate a paper that proposes a Bayesian nonparametric model for analyzing timeless multiple series. This study considered a P-order autoregressive structure for each series and borrowed power throughout the series, considering a common error population that also evolves over time.

The article explored prior properties and showed how to obtain posterior inferences. The model was tested in a simulation study and was illustrated by analyzing the economic activity index in the 32 states of Mexico.

For more information, visit our website.