A0605
Title: Dimension reduction for multivariate time series
Authors: Andreas Artemiou - University of Limassol (Cyprus) [presenting]
Abstract: The aim is to present two new methods for sufficient dimension reduction in multivariate time series. Following recent advances in the field of sufficient dimension reduction for multivariate time series, it is demonstrated how one can improve the performance of the existing methodologies by using a time series version of the sliced inverse mean difference and by improving the robustness to outliers by using a time series version of the sliced inverse median difference regression.
