Boundary-enhanced time series data imputation with long-term dependency diffusion models

TL;DR

This paper introduces a diffusion-based framework for multivariate time series data imputation that effectively handles boundary issues and captures long-term dependencies, outperforming existing methods.

Contribution

The paper proposes a novel diffusion model with boundary-aware strategies and a multi-scale S4-based U-Net to improve imputation accuracy in time series data.

Findings

Outperforms existing imputation methods in experiments

Effectively mitigates boundary inconsistencies

Captures long-term dependencies through multi-scale modeling

Abstract

Data imputation is crucial for addressing challenges posed by missing values in multivariate time series data across various fields, such as healthcare, traffic, and economics, and has garnered significant attention. Among various methods, diffusion model-based approaches show notable performance improvements. However, existing methods often cause disharmonious boundaries between missing and known regions and overlook long-range dependencies in missing data estimation, leading to suboptimal results. To address these issues, we propose a Diffusion-based time Series Data Imputation (DSDI) framework. We develop a weight-reducing injection strategy that incorporates the predicted values of missing points with reducing weights into the reverse diffusion process to mitigate boundary inconsistencies. Further, we introduce a multi-scale S4-based U-Net, which combines hierarchical information from different levels via multi-resolution integration to capture long-term dependencies. Experimental results demonstrate that our model outperforms existing imputation methods.