Causal Feature Selection for Weather-Driven Residential Load Forecasting

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Zhang, Elise, Mirallès, François, Dellacherie, Stéphane, Wu, Di, Boulet, Benoit
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915637546188800
author Zhang, Elise
Mirallès, François
Dellacherie, Stéphane
Wu, Di
Boulet, Benoit
author_facet Zhang, Elise
Mirallès, François
Dellacherie, Stéphane
Wu, Di
Boulet, Benoit
contents Weather is a dominant external driver of residential electricity demand, but adding many meteorological covariates can inflate model complexity and may even impair accuracy. Selecting appropriate exogenous features is non-trivial and calls for a principled selection framework, given the direct operational implications for day-to-day planning and reliability. This work investigates whether causal feature selection can retain the most informative weather drivers while improving parsimony and robustness for short-term load forecasting. We present a case study on Southern Ontario with two open-source datasets: (i) IESO hourly electricity consumption by Forward Sortation Areas; (ii) ERA5 weather reanalysis data. We compare different feature selection regimes (no feature selection, non-causal selection, PCMCI-causal selection) on city-level forecasting with three different time series forecasting models: GRU, TCN, PatchTST. In the feature analysis, non-causal selection prioritizes radiation and moisture variables that show correlational dependence, whereas PCMCI-causal selection emphasizes more direct thermal drivers and prunes the indirect covariates. We detail the evaluation pipeline and report diagnostics on prediction accuracy and extreme-weather robustness, positioning causal feature selection as a practical complement to modern forecasters when integrating weather into residential load forecasting.
format Preprint
id arxiv_https___arxiv_org_abs_2511_20508
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Causal Feature Selection for Weather-Driven Residential Load Forecasting
Zhang, Elise
Mirallès, François
Dellacherie, Stéphane
Wu, Di
Boulet, Benoit
Systems and Control
Weather is a dominant external driver of residential electricity demand, but adding many meteorological covariates can inflate model complexity and may even impair accuracy. Selecting appropriate exogenous features is non-trivial and calls for a principled selection framework, given the direct operational implications for day-to-day planning and reliability. This work investigates whether causal feature selection can retain the most informative weather drivers while improving parsimony and robustness for short-term load forecasting. We present a case study on Southern Ontario with two open-source datasets: (i) IESO hourly electricity consumption by Forward Sortation Areas; (ii) ERA5 weather reanalysis data. We compare different feature selection regimes (no feature selection, non-causal selection, PCMCI-causal selection) on city-level forecasting with three different time series forecasting models: GRU, TCN, PatchTST. In the feature analysis, non-causal selection prioritizes radiation and moisture variables that show correlational dependence, whereas PCMCI-causal selection emphasizes more direct thermal drivers and prunes the indirect covariates. We detail the evaluation pipeline and report diagnostics on prediction accuracy and extreme-weather robustness, positioning causal feature selection as a practical complement to modern forecasters when integrating weather into residential load forecasting.
title Causal Feature Selection for Weather-Driven Residential Load Forecasting
topic Systems and Control
url https://arxiv.org/abs/2511.20508