Multi-scale Self-Organizing Map assisted Deep Autoencoding Gaussian Mixture Model for unsupervised intrusion detection

作者：

Highlights：

• This study proposes SOM-DAGMM for unsupervised intrusion detection.

• This study further designs multi-scale SOM-DAGMMs to involve more SOM features.

• The proposed SOM-DAGMMs are able to better preserve the input space topology.

• This study designs and conducts extensive experiments to fully evaluate the performance of SOM-DAGMMs.

摘要

•This study proposes SOM-DAGMM for unsupervised intrusion detection.•This study further designs multi-scale SOM-DAGMMs to involve more SOM features.•The proposed SOM-DAGMMs are able to better preserve the input space topology.•This study designs and conducts extensive experiments to fully evaluate the performance of SOM-DAGMMs.

论文关键词：Intrusion detection,Anomaly detection,Self-Organizing Map,Input space topology,Deep Autoencoding Gaussian Mixture Model,Unsupervised training

论文评审过程：Received 28 December 2020, Revised 23 March 2021, Accepted 26 April 2021, Available online 27 April 2021, Version of Record 1 May 2021.

论文官网地址：https://doi.org/10.1016/j.knosys.2021.107086