AutoML Algorithms for Online Generalized Additive Model Selection: Application to Electricity Demand Forecasting

2025-03-31

AI Summary

Joint optimization of formula structure and state-space adaptive parameters in Generalized Additive Models (GAMs) remains challenging for short-term electricity load forecasting, particularly under online learning constraints. Method: We propose AutoGAM—the first automated GAM modeling framework tailored for online learning—by generalizing the neural architecture search method DRAGON to jointly optimize formula selection and adaptive parameterization within a differentiable, computationally efficient, and incrementally updatable search space. AutoGAM integrates state-space modeling with online joint hyperparameter optimization. Contribution/Results: Evaluated on real-world short-term electricity demand forecasting in France, AutoGAM achieves a 12.7% reduction in MAE over baseline methods while demonstrating superior dynamic adaptability, robustness, and low-latency inference. The framework eliminates manual design and offline tuning bottlenecks, establishing a new paradigm for automated, adaptive GAM deployment in time-critical operational settings.

Technology Category

• Machine Learning: Auto ML and Hyperparameter Tuning
• Search and Optimization: Learning to Search
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• User Modeling, Personalization and Recommendation: User modeling for targeted and personalized online advertising

Abstract

Electricity demand forecasting is key to ensuring that supply meets demand lest the grid would blackout. Reliable short-term forecasts may be obtained by combining a Generalized Additive Models (GAM) with a State-Space model (Obst et al., 2021), leading to an adaptive (or online) model. A GAM is an over-parameterized linear model defined by a formula and a state-space model involves hyperparameters. Both the formula and adaptation parameters have to be fixed before model training and have a huge impact on the model's predictive performance. We propose optimizing them using the DRAGON package of Keisler (2025), originally designed for neural architecture search. This work generalizes it for automated online generalized additive model selection by defining an efficient modeling of the search space (namely, the space of the GAM formulae and adaptation parameters). Its application to short-term French electricity demand forecasting demonstrates the relevance of the approach

Problem

• Optimizing GAM formula and parameters for electricity demand forecasting
• Adapting AutoML for online generalized additive model selection
• Improving predictive performance via efficient search space modeling

Innovation

• AutoML optimizes GAM and State-Space models
• DRAGON package enables automated model selection
• Efficient search space modeling for GAM formulae