VOS: A new outlier detection model using virtual graph

作者：

Highlights：

• We devise a virtual graph to represent the potential outlier-ness of each object within the dataset.

• A tailored Markov random walk process is applied to the virtual graph by utilizing both of the local and global information.

• The effectiveness of the proposed model is guaranteed by theoretical and experimental analyses.

摘要

•We devise a virtual graph to represent the potential outlier-ness of each object within the dataset.•A tailored Markov random walk process is applied to the virtual graph by utilizing both of the local and global information.•The effectiveness of the proposed model is guaranteed by theoretical and experimental analyses.

论文关键词：Anomaly detection,Outlier detection,Graph-based outlier detection,Neighborhood information graph,Virtual graph,Markov random walk

论文评审过程：Received 16 November 2018, Revised 30 July 2019, Accepted 1 August 2019, Available online 2 August 2019, Version of Record 25 October 2019.

论文官网地址：https://doi.org/10.1016/j.knosys.2019.104907