Autoregressive forests for multivariate time series modeling
作者：
Highlights：
• A novel autoregressive approach to represent multivariate time series is proposed.
• Proposed method is a non-linear, non-parametric vector autoregressive approach.
• Proposed method requires few parameters to tune and it is robust to the changes in parameters if they are set in a certain range.
• The proposed method can handle large number of features and long time series effectively.
• The approach provides competitive results in several multivariate time series classification problems.
摘要
•A novel autoregressive approach to represent multivariate time series is proposed.•Proposed method is a non-linear, non-parametric vector autoregressive approach.•Proposed method requires few parameters to tune and it is robust to the changes in parameters if they are set in a certain range.•The proposed method can handle large number of features and long time series effectively.•The approach provides competitive results in several multivariate time series classification problems.
论文关键词：Multivariate time series,Vector autoregression,Time series representation,Ensemble learning,Classification
论文评审过程：Received 13 March 2017, Revised 4 August 2017, Accepted 15 August 2017, Available online 24 August 2017, Version of Record 18 September 2017.
论文官网地址：https://doi.org/10.1016/j.patcog.2017.08.016
