O metodzie prognozowania brakujących danych w dziennych szeregach czasowych z lukami systematycznymi
About method of forecasting the missing data in daily time series with systematic gaps

Summary/Abstract: This paper presents a method for modeling and then forecasting in situation, when in time series for daily data contain systematic gaps. Base of construction were regular hierarchical time series models describing annual fluctuations. Weekly fluctuations were described as a grouping variable, which contains similar days and this type variables for other days. In models were used also dummy variables describing holidays and days pre- and post- holidays. Theoretical considerations were illustrated by empirical example for selected variant of gaps. Based on the same estimated equations, inter- and extrapolation predictions ware built. For both types of prediction – in general and in disaggregation to weekdays and months and holidays and days pre- and post holidays.

• Details
• Contents