DMS Colloquium: Roberto Molinari

Speaker: Roberto Molinari (Auburn University)

Title:  Robust Inference for Large Scale Stochastic Processes

Abstract: Performing robust statistical inference (i.e., reducing the impact of outliers) for different time series models has been an important challenge from a computational perspective. We put forward the Robust Generalized Method of Wavelet Moments in order to robustly estimate and perform inference in a computationally efficient manner even for more complex (state-space) models and for large data settings. This opens up possibilities in other areas of robust inference which also have been highly challenging in practice (e.g., model selection and spatial data).

Faculty host: Mark Carpenter