ARNISA SOKOLI, ERALDA GJIKA (DHAMO)
KEYWORDS : energy, market, seasonal, statistical analysis, demand, production.
Abstract
Prediction models play a critical role in understanding dynamic relationships within energy markets, providing insights into price fluctuations and trading volumes. By capturing these complex interactions, advanced forecasting techniques enable better decision-making and strategic planning. The establishment of the Albanian Power Exchange (ALPEX) has introduced a structured energy market in Albania, fostering transparency and competitiveness in electricity trading. This study focuses on modelling energy market dynamics in Albania by analysing key indicators, including Market Clearing Prices (MCP), energy volumes traded, and energy production and consumption. Using daily data from May 2023 to May 2024, the study employs both statistical and machine learning models to forecast MCP as a univariate time series. By comparing the predictive performance of traditional statistical approaches such as ARIMA with advanced machine learning techniques like XGBoost and Neural Network Autoregressive, the research aims to identify the most accurate methodologies for forecasting MCP. The findings aim to guide stakeholders in utilizing data-driven models for market analysis and strategic development.