Date of Publication :5th December 2017

Abstract: Security in corporations is a crucial issue. As number of users in these corporation increases, the chance for having intruder also increases. It is important to develop effective methods to deal with such threat. Luckily, users leave electric footprint behind, as log files. Analyzing these log files results in examining users’ activity and detecting an intruder. Recent works have proposed methods for detecting intruders inside corporations. However, these methods are complex for today’s corporation. In this work, we proposed a lightweight and effective method to detect an intruder inside corporations using log files. The dataset in this work was provided from NextLabs, one of high-profile companies in information security. The experiment using random forest algorithm shows that this method detects intruders with 97.18% accuracy.

Reference :