Applicability of time series combined models in forecasting the regional labour market indicators

The paper has a methodological character. It deals with possibilities of univariate time series models use in forecasting the regional labour market indicators. The main attention is focused on methodology for combining forecasts from different individual procedures. Forecasting performances of both individual and combined models are examined on a set of real time series of indicators of South Bohemia Region. The findings of an empirical comparative analysis demonstrate that forecast accuracy can be considerably improved through the combination of two or three individual forecasts.