| dc.description.abstract | This thesis uses electricity data sourced from Nord Pool and weather data obtained from Norsk Klimaservicesenter, seeking to forecast day-ahead spot prices by leveraging temperature-based demand forecasts. Through this analysis, we aim to examine the feasibility of developing a model that can be utilised by participants in the electricity market bidding process. A significant portion of our research efforts has been dedicated to exploring a SARIMAX model, which is widely employed in this field of research. However, we have also thoroughly examined and tested various alternative models to assess their viability by considering them as potential benchmarks.
The thesis is structured into several chapters, beginning with an initial introduction that provides an overview of the electricity market in Norway. This section serves to establish the context and background for our research. Following the introduction, we delve into the presentation of the data and methods employed to address our research question. This chapter outlines the specific datasets utilised and the methodologies implemented in our analysis. Finally, we conclude the thesis by presenting our results and the implications our study might have for the participants in the Nord Pool day-ahead market.
Our findings reveal a notable spurious correlation between temperature and spot price. However, we acknowledge that relying solely on weather variables is insufficient due to the influence of external factors on pricing decisions. Nevertheless, our research has yielded satisfactory results, with the best models achieving an overall error ranging between 5-10%. Our main model consistently performed well, although there were instances where alternative models outperformed it on specific days or weeks. | en_US |
