GMiner: Rule-Based Fuzzy Clustering for Google Drive Behavioral Type Mining
Chih-Hung Hsieh, Cheng-Hao Yan, Ching-Hao Mao, Chi-Ping Lai, Jenq‐Shiou Leu · 2016
Due to more and more on-premises services are migrating onto cloud, user behavioral analysis then gets popular as a data-driven way to administer lots accounts of on-cloud services. This paper proposes a novel rule-based approach, GMiner, for mining different types of Google cloud drive usages as an unsupervised account-management approach. Experiment results show that GMiner provides accurate, inter-pretable, and visualized clustering results which are helpful for highlighting inactive, quasi-insider accounts, or other potential cyber-security risks from real-environment dataset.