Efficient subspace search in data streams

作者：

Highlights：

• A general framework, SGMRD, to monitor subspaces in high-dimensional data streams.

• SGMRD leverages multivariate dependency estimators and bandit algorithms.

• SGMRD is efficient and helps with downstream mining tasks, e.g., outlier detection.

• The experiments show the superiority of SGMRD w.r.t. the existing work.

摘要

•A general framework, SGMRD, to monitor subspaces in high-dimensional data streams.•SGMRD leverages multivariate dependency estimators and bandit algorithms.•SGMRD is efficient and helps with downstream mining tasks, e.g., outlier detection.•The experiments show the superiority of SGMRD w.r.t. the existing work.

论文关键词：Subspace search,Data stream monitoring,Outlier detection

论文评审过程：Received 16 November 2020, Revised 10 December 2020, Accepted 11 December 2020, Available online 17 December 2020, Version of Record 30 December 2020.

论文官网地址：https://doi.org/10.1016/j.is.2020.101705