Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep Learning for Multivariate Time Series Imputation: A Survey

Jun Wang, Wenjie Du, Yiyuan Yang, Linglong Qian, Wei Cao, Keli Zhang, Wenjia Wang, Yuxuan Liang, Qingsong Wen

TL;DR

This survey addresses the pervasive problem of missing values in multivariate time series by classifying deep learning approaches along two axes: imputation uncertainty (predictive vs generative) and neural architecture (RNN, CNN, GNN, attention, VAE, GAN, diffusion, PFMs, LLMs). It systematically reviews predictive and generative methods, discusses a range of architectures, and covers large-model trends and time series imputation toolkits like PyPOTS and TSI-Bench. The authors outline practical challenges, including MNAR missingness, downstream task integration, and scalability, and propose future directions involving foundation models, multimodal data, and efficient diffusion or transformer-based methods. Overall, the work serves as a comprehensive reference for researchers and practitioners seeking to advance robust, uncertainty-aware MTS imputation in real-world settings.

Abstract

Missing values are ubiquitous in multivariate time series (MTS) data, posing significant challenges for accurate analysis and downstream applications. In recent years, deep learning-based methods have successfully handled missing data by leveraging complex temporal dependencies and learned data distributions. In this survey, we provide a comprehensive summary of deep learning approaches for multivariate time series imputation (MTSI) tasks. We propose a novel taxonomy that categorizes existing methods based on two key perspectives: imputation uncertainty and neural network architecture. Furthermore, we summarize existing MTSI toolkits with a particular emphasis on the PyPOTS Ecosystem, which provides an integrated and standardized foundation for MTSI research. Finally, we discuss key challenges and future research directions, which give insight for further MTSI research. This survey aims to serve as a valuable resource for researchers and practitioners in the field of time series analysis and missing data imputation tasks.A well-maintained MTSI paper and tool list are available at https://github.com/WenjieDu/Awesome_Imputation.

Deep Learning for Multivariate Time Series Imputation: A Survey

TL;DR

This survey addresses the pervasive problem of missing values in multivariate time series by classifying deep learning approaches along two axes: imputation uncertainty (predictive vs generative) and neural architecture (RNN, CNN, GNN, attention, VAE, GAN, diffusion, PFMs, LLMs). It systematically reviews predictive and generative methods, discusses a range of architectures, and covers large-model trends and time series imputation toolkits like PyPOTS and TSI-Bench. The authors outline practical challenges, including MNAR missingness, downstream task integration, and scalability, and propose future directions involving foundation models, multimodal data, and efficient diffusion or transformer-based methods. Overall, the work serves as a comprehensive reference for researchers and practitioners seeking to advance robust, uncertainty-aware MTS imputation in real-world settings.

Abstract

Missing values are ubiquitous in multivariate time series (MTS) data, posing significant challenges for accurate analysis and downstream applications. In recent years, deep learning-based methods have successfully handled missing data by leveraging complex temporal dependencies and learned data distributions. In this survey, we provide a comprehensive summary of deep learning approaches for multivariate time series imputation (MTSI) tasks. We propose a novel taxonomy that categorizes existing methods based on two key perspectives: imputation uncertainty and neural network architecture. Furthermore, we summarize existing MTSI toolkits with a particular emphasis on the PyPOTS Ecosystem, which provides an integrated and standardized foundation for MTSI research. Finally, we discuss key challenges and future research directions, which give insight for further MTSI research. This survey aims to serve as a valuable resource for researchers and practitioners in the field of time series analysis and missing data imputation tasks.A well-maintained MTSI paper and tool list are available at https://github.com/WenjieDu/Awesome_Imputation.
Paper Structure (29 sections, 4 equations, 2 figures, 1 table)

This paper contains 29 sections, 4 equations, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Preliminary and Taxonomy
  3. Background of MTSI
  4. Problem Definition
  5. Missing Mechanism
  6. Taxonomy of Deep Learning-based MTSI
  7. Predictive Methods
  8. Learning Objective
  9. RNN-based Models
  10. CNN-based Models
  11. GNN-based Models
  12. Attention-based Models
  13. Pros and Cons
  14. Generative Methods
  15. Learning Objective
  16. ...and 14 more sections

Figures (2)

  • Figure 1: The framework of our survey.
  • Figure 2: The taxonomy of deep learning methods for multivariate time series imputation from the view of imputation uncertainty and neural network architecture.