Efficient classification using parallel and scalable compressed model and its application on intrusion detection

作者：

Highlights：

• We propose a compressed model composed of horizontal and vertical compression.

• We employ OneR as horizontal compression, AP clustering as vertical compression.

• We implement a Map-Reduce based scalable and parallel framework for compression.

• We use KNN and SVM to build intrusion detector using the proposed compressed model.

• Both KDD99 and CDMC2012 are evaluated to show the detection efficiency and accuracy.

摘要

•We propose a compressed model composed of horizontal and vertical compression.•We employ OneR as horizontal compression, AP clustering as vertical compression.•We implement a Map-Reduce based scalable and parallel framework for compression.•We use KNN and SVM to build intrusion detector using the proposed compressed model.•Both KDD99 and CDMC2012 are evaluated to show the detection efficiency and accuracy.

论文关键词：Compressed model,MapReduce,Parallelization,Classification,Intrusion detection

论文评审过程：Available online 13 April 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.04.009