Time series cluster kernel for learning similarities between multivariate time series with missing data
作者：
Highlights：
• The time series cluster kernel (TCK) for multivariate time series (MTS) is proposed.
• Gaussian mixture model (GMM) ensemble learning for increased parameter robustness.
• Robustness to missing data is ensured by extending the GMMs using informative priors.
• We prove that the TCK is a valid kernel.
• TCK outperforms established methods on missing data problems.
摘要
•The time series cluster kernel (TCK) for multivariate time series (MTS) is proposed.•Gaussian mixture model (GMM) ensemble learning for increased parameter robustness.•Robustness to missing data is ensured by extending the GMMs using informative priors.•We prove that the TCK is a valid kernel.•TCK outperforms established methods on missing data problems.
论文关键词：Multivariate time series,Similarity measures,Kernel methods,Missing data,Gaussian mixture models,Ensemble learning
论文评审过程：Received 23 March 2017, Revised 11 October 2017, Accepted 30 November 2017, Available online 6 December 2017, Version of Record 21 December 2017.
论文官网地址：https://doi.org/10.1016/j.patcog.2017.11.030
