Shape-based outlier detection in multivariate functional data

作者：

Highlights：

• Functional curve shape features are proposed to detect outliers in multivariate functional data.

• These functional features are combined with state-of-the-art outlier detection algorithms.

• Our proposal experimentally outperforms baseline methods in various outlier detection scenarios.

摘要

•Functional curve shape features are proposed to detect outliers in multivariate functional data.•These functional features are combined with state-of-the-art outlier detection algorithms.•Our proposal experimentally outperforms baseline methods in various outlier detection scenarios.

论文关键词：Outlier detection,Anomaly detection,Multivariate functional data,functional data features,Multivariate time series,Multidimensional curve shape

论文评审过程：Received 19 July 2019, Revised 27 March 2020, Accepted 20 April 2020, Available online 23 April 2020, Version of Record 30 April 2020.

论文官网地址：https://doi.org/10.1016/j.knosys.2020.105960