Time series imputation improves with global dataset knowledge integration
ProCTI: Prototype-Refined Global Conditioning for Diffusion-Based Time Series Imputation
Machine LearningArtificial Intelligence
Summary
Time series data often have missing gaps that need to be filled accurately. The authors show that existing methods mostly use nearby data points to guess missing values, which can fail when data is sparse or noisy. They introduce ProCTI, which combines local information with learned global patterns from the entire dataset. This helps fill in missing data better, especially when local information alone is unreliable.
What this means in practice
- •For data engineers: Improve accuracy of missing data restoration in time series from sensors or logs using global dataset patterns along with local context.
- •For financial analysts: Fill in missing financial time series data more reliably when some local observations are sparse or noisy by leveraging global market behavior patterns.
Authors
Fariza Rashid, Duc Van Le, Rahat Masood, Gustavo Batista, Aruna Seneviratne, Suranga Seneviratne
Abstract
Time series imputation has progressed from statistical and deep learning approaches to diffusion-based models, which have shown strong recent performance. Existing diffusion-based methods typically condition the reverse process using local contextual information from the current or neighbouring windows. Meanwhile, global dataset-level structure often remains implicit, limiting performance when local observations are sparse, noisy, or unrepresentative. To address this issue, we propose ProCTI, a diffusion-imputation framework that augments local conditioning with retrieved global dataset-level priors through learned prototypes. A hybrid conditioning mechanism integrates this global context with local signals during reverse diffusion, enabling more accurate reconstruction under varying missingness scenarios. Experiments across multiple benchmark datasets show that ProCTI outperforms strong baselines overall under random missingness, while remaining competitive under attribute-wise missingness. Furthermore, we use a latent-regime data model to characterise the precise conditions under which prototype-derived global conditioning provably improves imputation. We support this with a general theoretical analysis of local-global conditioning.