Asynchronism-based principal component analysis for time series data mining

作者：

Highlights：

• The proposed method APCA can process the time series of different length.

• A synchronous or an asynchronous correlation can be reflected by APCA.

• APCA retains much more information than PCA in the reduced dimension.

摘要

•The proposed method APCA can process the time series of different length.•A synchronous or an asynchronous correlation can be reflected by APCA.•APCA retains much more information than PCA in the reduced dimension.

论文关键词：Asynchronous correlation,Covariance matrix,Principal component analysis,Time series data mining,Dynamic time warping

论文评审过程：Available online 19 October 2013.

论文官网地址：https://doi.org/10.1016/j.eswa.2013.10.019