Univariate time series classification using information geometry

作者：

Highlights：

• Scalar time series is unfolded to phase space to find its original structure.

• Covariance matrix fuses global features, local features and their interactions.

• Classification is carried out in the tangent space of statistical manifolds.

摘要

•Scalar time series is unfolded to phase space to find its original structure.•Covariance matrix fuses global features, local features and their interactions.•Classification is carried out in the tangent space of statistical manifolds.

论文关键词：Time series,Classification,Information geometry,Riemannian manifold

论文评审过程：Received 18 December 2018, Revised 17 March 2019, Accepted 30 May 2019, Available online 31 May 2019, Version of Record 5 June 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.05.040