Machine Learning for Electricity Price Forecasting in Serbia

Created using ChatSlide

This project aims to forecast day-ahead electricity prices in Serbia, a market facing unique challenges. The research surveys forecasting methods, highlighting limited studies on Southeast Europe, and focusses on using SARIMA and XGBoost models to predict prices using market, demand, and weather data. A comprehensive analysis covers Serbia's energy dynamics, the role of seasonality, and data peculiarities. Model development prioritises capturing seasonality with SARIMA and managing dynamic...