Practical decision-making in electricity consumption forecasting

: insights from combined models

Abstract

This study focuses on forecasting daily electricity consumption at Faro Airport, utilizing machine learning models such as Random Forest (RF) and AutoRegressive Integrated Moving Average with exogenous variables (ARIMAX). In addition to demonstrating the forecasting prowess of these models, this study evaluates their performance by comparing them with the ensembled model blending RF and ARIMAX. Performance metrics not only gauge the accuracy of predictions but also provide a nuanced understanding of the models’ effectiveness. The hybrid model emerges as a standout performer, showcasing superior forecasting precision. Its ability to leverage the strengths of both RF and ARIMAX contributes to more robust predictions, especially in the context of daily electricity consumption at Faro Airport. Beyond numerical accuracy, the study incorporates Shapley values for interpretability, offering a transparent view of the factors influencing electricity consumption trends. This interpretability aids airport management in making informed decisions related to en ergy resource allocation, infrastructure planning, and operational efficiency. This approach not only optimizes resource utilization but also positions the airport to proactively address challenges and opportunities in its energy consumption patterns. The combination of accurate predictions and interpretable insights positions Faro Airport to not only meet current energy demands effectively but also to navigate future challenges with strategic foresight.

Date of Award	30 Jan 2024
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Ana Marisa Mendes Gonçalves Vinhais Guedes (Supervisor)