Presentations at an Academic or Professional conference
Comparative Study of Modelling Time Series with Outliers
The aim of this paper is to present and compare various types of modeling of the impact of outliers on time series. We use ARIMA models with outliers (AO, IO, LS and TC) and exponential smoothing. The data studied concern food products subjected to promotional marketing and the effects on air traffic of the existence of a new kind of transport : the HST (French TGV).
INDJEHAGOPIAN, J.P. et MACE, S. (1994). Comparative Study of Modelling Time Series with Outliers.