A method for recovery of multidimensional time series based on the detection of behavioral patterns and the use of autoencoders

This article presents a method for recovering missing values in multidimensional time series. The method combines neural network technologies and an algorithm for searching snippets (behavioral patterns of a time series). It includes the stages of data preprocessing, recognition and reconstruction, using convolutional and recurrent neural networks. Experiments have shown high accuracy of recovery and the advantage of the method over SOTA methods.

cs.AI Computer Science

View graph Source

Paper

References (9)

01On the imputation of missing data for road traffic forecasting: New insights and novel techniquesI. Laña, I. Olabarrieta, M. Vélez et al.2018 · 99 citations

02Estimating Missing Data in Temporal Data Streams Using Multi-Directional Recurrent Neural NetworksJinsung Yoon, W. Zame, M. van der Schaar2017 · IEEE Transactions on Biomedical Engineering · 345 citations In Library

03Conditional Variational Autoencoder for Prediction and Feature Recovery Applied to Intrusion Detection in IoTManuel López Martín, B. Carro, Antonio J. Sánchez-Esguevillas et al.2017 · Italian National Conference on Sensors · 267 citations

04Robust Orthonormal Subspace Learning: Efficient Recovery of Corrupted Low-Rank MatricesXianbiao Shu, F. Porikli, N. Ahuja2014 · 2014 IEEE Conference on Computer Vision and Pattern Recognition · 97 citations

05DynaMMo: mining and summarization of coevolving sequences with missing valuesLei Li, J. Mccann, N. Pollard et al.2009 · Knowledge Discovery and Data Mining · 174 citations

06A Singular Value Thresholding Algorithm for Matrix CompletionJian-Feng Cai, E. Candès, Zuowei Shen2008 · SIAM Journal on Optimization · 6.1k citations

07Mind the Gap: An Experimental Evaluation of Imputation of Missing Values Techniques in Time SeriesMourad Khayati, Alberto Lerner, Zakhar Tymchenko2020 · 75 citations

08: Deep probabilistic time series imputation2020 · International conference on artificial intelligence and statistics

09Dying ReLU and Initialization: Theory and Numerical Examples2020 · Commun. Comput. Phys.

Similar papers

Time Series Anomaly Detection System with Linear Neural Network and AutoencoderPritika Mehra, Mini Singh Ahuja, Manisha Aeri2023 In Library

Pattern discovery in time series using autoencoder in comparison to nonlearning approachesF. Noering, Yannik Schröder, K. Jonas et al.2021 · Integr. Comput. Aided Eng. · 16 citations In Library

Learning representations for multivariate time series with missing data using Temporal Kernelized AutoencodersF. Bianchi, L. Livi, Karl Øyvind Mikalsen et al.2018 · arXiv.org · 11 citations In Library

Multivariate Time Series as Images: Imputation Using Convolutional Denoising AutoencoderA. Safi, C. Beyer, Vishnu Unnikrishnan et al.2020 · International Symposium on Intelligent Data Analysis · 9 citations In Library

Multivariate Time Series Missing Data Imputation Using Recurrent Denoising AutoencoderJianye Zhang, Peng Yin2019 · IEEE International Conference on Bioinformatics and Biomedicine · 36 citations In Library

Memory-augmented Adversarial Autoencoders for Multivariate Time-series Anomaly Detection with Deep Reconstruction and PredictionQinfeng Xiao, Shikuan Shao, Jing Wang2021 · arXiv (Cornell University) · 10 citations In Library

Anomaly Detection in Multivariate Time Series Data Using Auto encoder and LSTM Networks: A Hybrid Deep Learning ApproachProf. Smarika Rai2026 · Zenodo (CERN European Organization for Nuclear Research) In Library

Anomaly Detection in Multivariate Time Series Data Using Auto encoder and LSTM Networks: A Hybrid Deep Learning ApproachProf. Smarika Rai2026 · Zenodo (CERN European Organization for Nuclear Research) In Library

Using Autoencoders and AutoDiff to Reconstruct Missing Variables in a Set of Time SeriesJan-Philipp Roche, Oliver Niggemann, Jens Friebe2023 · arXiv (Cornell University) · 3 citations In Library

Enhancing Reconstruction based Multivariate Time Series Anomaly Detection via Reinforcement LearningRenlong Zhang, Yu Xu, Yiheng Lu et al.2025 In Library

Source of record

arXiv

44B

Welcome.

This panel sticks with you. Pick anything below and it opens right beside it, so you can dig through 60,000-plus records without ever losing your spot here.

Travel 44B

LibraryPapers, policy, standards, statute LabsEvery organization building AI ModelsIntelligence, price, and speed BenchmarksThe evaluation catalog IncidentsAI failures and harms SearchOne field across everything DashboardThe state of AI accountability in NY 44B RegistryThe Art. 44-B compliance portal