Variational auto-encoders based on the shift correction for imputation of specific missing in multivariate time series

被引:16
作者
Li, Junying [1 ]
Ren, Weijie [2 ]
Han, Min [3 ,4 ]
机构
[1] Dalian Univ Technol, Fac Elect Informat & Elect Engn, Dalian 116024, Peoples R China
[2] Harbin Engn Univ, Coll Intelligent Syst Sci & Engn, Harbin 150001, Peoples R China
[3] Dalian Univ Technol, Key Lab Intelligent Control & Optimizat Ind Equip, Minist Educ, Dalian 116024, Peoples R China
[4] Dalian Univ Technol, Chinad Profess Technol Innovat Ctr Distributed Co, Dalian 116024, Peoples R China
基金
中国国家自然科学基金;
关键词
Missing data; Imputation; Variational auto-encoders; Shift correction; Multivariate time series;
D O I
10.1016/j.measurement.2021.110055
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
Data missing is a ubiquitous phenomenon in multivariate time series analysis because of failure measurements, improper installation or other unfavorable factors. Most imputation methods, such as statistical methods and machine learning methods, are mainly for dealing with random or continuous missing, but do not take into consideration the non-random specific missing situation. This paper proposes an imputation model based on the variational auto-encoders (VAE) and shift correction for specific missing values, which is also extended to the beta-VAE model. The shift correction is used to correct the original probability distribution deviation caused by specific values concentrated missing. Then taking the meteorological and air quality data sets in Beijing as the example, this paper mainly explores two common specific missing situations, including large values missing and small values missing. The experimental results show that the proposed model can effectively impute specific missing values, which has high imputation accuracy and robustness.
引用
收藏
页数:15
相关论文
共 51 条
[1]   Missing data imputation using fuzzy-rough methods [J].
Amiri, Mehran ;
Jensen, Richard .
NEUROCOMPUTING, 2016, 205 :152-164
[2]  
Arjovsky M., 2017, ARXIV17010785V3
[3]  
Beaulieu-Jones BK, 2017, BIOCOMPUT-PAC SYM, P207, DOI 10.1142/9789813207813_0021
[4]  
Biessmann F, 2019, J MACH LEARN RES, V20
[5]   A variational autoencoder solution for road traffic forecasting systems: Missing data imputation, dimension reduction, model selection and anomaly detection [J].
Boquet, Guillem ;
Morell, Antoni ;
Serrano, Javier ;
Lopez Vicario, Jose .
TRANSPORTATION RESEARCH PART C-EMERGING TECHNOLOGIES, 2020, 115
[6]  
Boquet G, 2019, INT CONF ACOUST SPEE, P2882, DOI [10.1109/ICASSP.2019.8683011, 10.1109/icassp.2019.8683011]
[7]  
Brock A., 2016, ICLR
[8]  
Christopher P., 2018, NIPS WORKSH LEARN DI
[9]   Missing Data Imputation via Denoising Autoencoders: The Untold Story [J].
Costa, Adriana Fonseca ;
Santos, Miriam Seoane ;
Soares, Jastin Pompeu ;
Abreu, Pedro Henriques .
ADVANCES IN INTELLIGENT DATA ANALYSIS XVII, IDA 2018, 2018, 11191 :87-98
[10]  
Dai B, 2018, The Journal of Machine Learning Research, V19, P1573