Title: A X11-based seasonal adjustment method for series with multiple periodicities
Authors: Dominique Ladiray - INSEE (France) [presenting]
Abstract: The increasing availability of more and higher frequency data has opened new challenging ways of producing official statistics. More specifically, the question has arisen how established statistical techniques, such as seasonal adjustment, can be applied to large datasets of high frequency data and data with high periodicity (infra-monthly and infra-weekly periodicities). Even if the X-13ARIMA-SEATS method focuses on monthly and quarterly series, we show how the X11-algorithm can be generalized to multiple seasonalities. We also provide alternatives to the Reg-ARIMA feature used by X-13ARIMA-SEATS to automatically detect outliers and calendar effects and to forecast the series. Several examples are presented to illustrate the methodology using daily demography series (businesses and persons) and electricity demand.