Pattern classification-based electricity price interval forecasting with discrete–continuous inputs

The worldwide restructuring of electricity markets toward deregulation over the past few decades has attracted significant research attention to electricity price forecasting due to its importance for market participants. In recent years, market participants have been more concerned with forecasting the general price level than the exact price value, as this better supports bidding strategies and investment decision-making. However, current interval forecasting approaches face key challenges in efficiency, robustness against volatility, and ineffective periodicity extraction techniques. To address the limitations of existing electricity market price forecasting methods and to develop more practical strategies for real-world market conditions, we propose a novel model that integrates TimesBlock, an information fusion module designed to process multi-periodicity, with long short-term memory (LSTM) networks for long-term temporal modeling. This integration enables accurate pattern classification-based interval forecasting of electricity prices. An innovative discrete–continuous hybrid data input strategy is designed to both enhance the model’s ability to handle volatile data while preserving essential periodic information, and support lightweight deployment suitable for real-time applications—an aspect largely overlooked by existing studies. The experimental results demonstrate that our model provides reliable electricity price interval forecasts, outperforms other baselines on multiple metrics, and achieves significantly faster running speed due to the dual-mode input strategy. • A novel TimesBlock-LSTM framework is developed for electricity price interval forecasting. • An improved FFT-based extraction method captures more representative periodic patterns. • A discrete–continuous input strategy enhances model robustness and training speed. • Comprehensive studies verify model’s effectiveness and demonstrate superior performance.