Multidimensional surrogate stability to detect data stream concept drift
作者：
Highlights：
• We present an approach that detects drifts on multidimensional streams.
• The detection occurs while holding a stability property based on surrogate series.
• Our approach employs a different and more robust measurement to analyze drifts.
• MDFT allows analyze data dependencies of unidimensional streams in phase space.
• Experiments confirmed MDFT outperforms PHT, ADWIN and CUSUM.
摘要
•We present an approach that detects drifts on multidimensional streams.•The detection occurs while holding a stability property based on surrogate series.•Our approach employs a different and more robust measurement to analyze drifts.•MDFT allows analyze data dependencies of unidimensional streams in phase space.•Experiments confirmed MDFT outperforms PHT, ADWIN and CUSUM.
论文关键词：Data streams,Concept drift,Multidimensional,Surrogate data
论文评审过程：Received 17 August 2016, Revised 27 May 2017, Accepted 2 June 2017, Available online 7 June 2017, Version of Record 17 July 2017.
论文官网地址：https://doi.org/10.1016/j.eswa.2017.06.005
