Abstract:
We consider outlier detection algorithms for time series regression based on iterated 1-step Huber-skip M-estimators. This paper analyses the role of varying cut-offs in such algorithms. The argument involves an asymptotic theory for a new class of weighted and marked empirical processes allowing for estimation errors of the scale and the regression coefficient.
Citation:
Jiao, X., Nielsen, B. (2017), 'Asymptotic Analysis of Iterated 1-Step Huber-Skip M-Estimators with Varying Cut-Offs' In: Antoch, J., Jurečková, J., Maciak, M., Pešta, M. (eds) Analytical Methods in Statistics, AMISTAT 2015, Springer Proceedings in Mathematics & Statistics, vol 193. Springer, Cham. https://doi.org/10.1007/978-3-319-51313-3_2