Seasonal variation indicators calculated on the basis of empirical time series values and values of the series smoothed by some numerical method (e.g. moving average). The values of the empirical series are divided by the corresponding values of the smoothed series. The obtained quotients represent periodical (seasonal) variations and random variations. Random variations can be eliminated from the seasonality indicators by computing rough seasonal variation indicators.
GUS – Departament Programowania, Koordynacji Badań i Rejestrów