Dual-Splitting Conformal Prediction for Multi-Step Time Series Forecasting

2025-03-27

AI Summary

Existing conformal prediction (CP) methods for multi-step time series forecasting face limitations in quantifying uncertainty due to restrictive single-step assumptions, reliance on real-time data, and poor scalability. To address this, we propose the Dual-Split Conformal Prediction (DSCP) framework—a model-agnostic, two-stage data splitting strategy that jointly incorporates temporal dependency modeling and quantile regression. DSCP introduces dependency-aware calibration and constructs adaptive confidence sets via a rolling-window scheme, enabling statistically valid multi-step interval forecasts. Evaluated on four real-world datasets across diverse domains, DSCP achieves up to a 23.59% improvement in Winkler score over baselines. Applied to energy and IT workload forecasting for data center operations, it enables proactive resource optimization and reduces carbon emissions by 11.25%.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Sustainability and carbon-aware systems for Web, mobile, and WoT
• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Time series forecasting is crucial for applications like resource scheduling and risk management, where multi-step predictions provide a comprehensive view of future trends. Uncertainty Quantification (UQ) is a mainstream approach for addressing forecasting uncertainties, with Conformal Prediction (CP) gaining attention due to its model-agnostic nature and statistical guarantees. However, most variants of CP are designed for single-step predictions and face challenges in multi-step scenarios, such as reliance on real-time data and limited scalability. This highlights the need for CP methods specifically tailored to multi-step forecasting. We propose the Dual-Splitting Conformal Prediction (DSCP) method, a novel CP approach designed to capture inherent dependencies within time-series data for multi-step forecasting. Experimental results on real-world datasets from four different domains demonstrate that the proposed DSCP significantly outperforms existing CP variants in terms of the Winkler Score, achieving a performance improvement of up to 23.59% compared to state-of-the-art methods. Furthermore, we deployed the DSCP approach for renewable energy generation and IT load forecasting in power management of a real-world trajectory-based application, achieving an 11.25% reduction in carbon emissions through predictive optimization of data center operations and controls.

Problem

• Addresses multi-step time series forecasting challenges
• Improves uncertainty quantification in multi-step scenarios
• Enhances renewable energy and IT load forecasting

Innovation

• Dual-Splitting Conformal Prediction for multi-step forecasting
• Captures time-series dependencies with statistical guarantees
• Reduces carbon emissions via predictive optimization