Time Series Robust Methods - In the Time and Frequency Domains

Speakers: Valdério Anselmo Reisen, Pascal Bondon

This seminar aims to discuss recent developments in robust estimation in linear time series models, with short and long memory correlation structures, in the presence of additive outliers. Robust estimators of the auto covariance matrix will be discussed from both time and domain approaches from both theoretical and applied points of view. A variety of application models will be considered for the use of the proposed methodologies, such as multivariate techniques (factor analysis and PCA), time series and mixed models. Real applications will also be discussed.