Evaluating time series similarity using concept-based models
作者：
Highlights：
• We propose using local concept-based models to describe individual time series.
• A global concept-based model can be used to describe an entire time series dataset.
• Concepts can be labeled, making the model straightforward to use for a human being.
• We evaluate similarity of time series by evaluating the similarity of local models.
• Proposed scheme can be applied as a backbone for clustering and classification.
摘要
•We propose using local concept-based models to describe individual time series.•A global concept-based model can be used to describe an entire time series dataset.•Concepts can be labeled, making the model straightforward to use for a human being.•We evaluate similarity of time series by evaluating the similarity of local models.•Proposed scheme can be applied as a backbone for clustering and classification.
论文关键词：Time series,Concept-based model,Similarity,Time series clustering,Time series classification,Fuzzy models
论文评审过程：Received 2 October 2020, Revised 24 September 2021, Accepted 23 November 2021, Available online 17 December 2021, Version of Record 27 December 2021.
论文官网地址：https://doi.org/10.1016/j.knosys.2021.107811
