LoadKAN: Interpretable Kolmogorov-Arnold Network for Electricity Load Forecasting

This study introduces LoadKAN, a novel hybrid framework that combines a feature-isolated temporal attention mechanism with a Kolmogorov-Arnold network (KAN) to address the lack of interpretability in deep learning-based electricity load forecasting.

• LoadKAN extracts temporal dynamics from each input feature independently before passing distilled representations to the KAN module for prediction.
• The model was evaluated on datasets from three representative U.S. electricity markets, remaining highly competitive against extensively-tuned state-of-the-art black-box benchmarks.
• Quantitative sensitivity analyses using KAN-learned activation functions revealed complex, market-specific dependencies between six distinct mobility patterns and electricity load.

LoadKAN enables granular analysis of learned non-linear relationships, generating insights into feature dependencies that are typically obscured by opaque black-box neural forecasting models.