Balanced Iterative Reducing and Clustering Using Hierarchies with Principal Component Analysis (PBirch) for Intrusion Detection over Big Data in Mobile Cloud Environment

Abstract

With the development of big data, mobile cloud computing, cyber security issues have become more and more critical. Thus, enabling an intrusion detection method over big data in mobile cloud environment is of paramount importance. In our previous research, we proposed an approach named Mini Batch Kmeans with Principal Component Analysis (PMBKM) for big data which can effectively solve the clustering problem for intrusion detection of big data, but it needs to preset the number of clusters. The best clustering number is selected by comparing the clustering results of different clustering values multiple times. To address the above issue, we propose a new clustering method named Balanced Iterative Reducing and Clustering Using Hierarchies with Principal Component Analysis (PBirch) in this paper. Compared to PMBKM, the experimental results show that PBirch can obtain a good clustering result without presetting clustering values, and the clustering result can be further improved by optimizing the relevant parameters. The clustering time of PBirch decreases linearly with the increasing of the cluster numbers. Thus, the larger the number of clusters, the smaller the PBirch time cost. All in all, our proposed method can be widely used for big data in mobile cloud environment.