Published Paper Details:

Recent studies have highlighted that insider threats are more destructive than external network threats. This project mainly focuses on the user behavior to detect the insider attack within the organization. Based on the above considerations, we have come up with some solutions where we focus on the behavior of individual user, User activities and analysis on the access rights and usage to check the outcome as whether they are Normal user with no harm or Abnormal (Malicious). Feature Engineering is the process where we select the fixed set of procedures to identify behavior of the employee effectively. The implementation is done by applying various machine learning and deep learning algorithms to get high classification accuracy of the model. The trained data is feed to the Model engine to gain the experience about the user activities and test data is used to find the accuracy of the model and defining the behavior of the user a normal or malicious. The data used is the CMU CERT synthetic insider threat dataset version r5.2 Our unique approach produces comparatively good accuracy of 100%.