Applied short-term forecasting for the Slovenian natural gas market

Results of applied short-term forecasting for the Slovenian natural gas market are presented. A case study for one of the major Slovenian natural gas distribution companies is considered, with forecasting results in hourly resolution in the forecasting horizon from 1 to 48 hours. The development of a forecasting strategy is presented, which includes daily data acquisition from various sources, development of customized forecasting models, and deployment of a forecasting solution. The forecasting models are based on stepwise regression method in order to design model structures with reliable and robust operation. Forecasting results of several years of online forecasting operation are presented. Obtained accuracy of forecasting results is considered as very successful and beneficially contributes to the managing policy of the distribution company.