The Living Forecast: Evolving Day-Ahead Predictions into Intraday Reality

2025-10-14

AI Summary

To address the degradation of day-ahead probabilistic forecasting accuracy over time due to information lag, this paper proposes a retraining-free Bayesian updating mechanism that dynamically refines day-ahead forecasts into intraday forecasts. Methodologically, it employs a conditional variational autoencoder to generate Gaussian mixture outputs and constructs covariance structure via a pattern dictionary; subsequent conditional updating of the predictive distribution is performed using real-time observations—preserving the original probabilistic structure while jointly producing point forecasts, quantile forecasts, and ensemble forecasts. The approach is theoretically rigorous, computationally efficient, and suitable for real-time deployment. Experiments on residential load and photovoltaic generation data demonstrate up to a 25% reduction in prediction error, with particularly pronounced improvements during periods of strong observational correlation.

Technology Category

• Machine Learning: Bayesian Learning
• Reasoning under Uncertainty: Relational Probabilistic Models
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Accurate intraday forecasts are essential for power system operations, complementing day-ahead forecasts that gradually lose relevance as new information becomes available. This paper introduces a Bayesian updating mechanism that converts fully probabilistic day-ahead forecasts into intraday forecasts without retraining or re-inference. The approach conditions the Gaussian mixture output of a conditional variational autoencoder-based forecaster on observed measurements, yielding an updated distribution for the remaining horizon that preserves its probabilistic structure. This enables consistent point, quantile, and ensemble forecasts while remaining computationally efficient and suitable for real-time applications. Experiments on household electricity consumption and photovoltaic generation datasets demonstrate that the proposed method improves forecast accuracy up to 25% across likelihood-, sample-, quantile-, and point-based metrics. The largest gains occur in time steps with strong temporal correlation to observed data, and the use of pattern dictionary-based covariance structures further enhances performance. The results highlight a theoretically grounded framework for intraday forecasting in modern power systems.

Problem

• Converting day-ahead probabilistic forecasts into updated intraday predictions
• Improving forecast accuracy for electricity consumption and photovoltaic generation
• Enabling real-time probabilistic updates without model retraining or re-inference

Innovation

• Bayesian updating mechanism converts probabilistic forecasts
• Gaussian mixture output conditioned on observed measurements
• Pattern dictionary-based covariance structures enhance performance