Seasonal dynamics in the prediction of household-level energy poverty: A machine learning approach

Key Points

• The aim is to predict seasonal energy poverty at the household level using machine learning techniques.
• Analyzed 330,960 household-month data points
• Employed machine learning approaches for prediction
• Investigated the impact of structural and socioeconomic factors on energy poverty
• Identified heating-related structural vulnerability as a key driver of energy poverty in winter
• Found that summer energy poverty is influenced by socioeconomic factors
• Double risk groups experience high vulnerability throughout both summer and winter seasons