Detecting Compromised User Account Based on User Behavior Analysis

Authors

  • Naeem Yousir

Abstract

Most researchers indicate the user is a major vector of an attacker. Currently, there is no complete description of the attacker scenario to alert the user on what information has been accessed too. In this paper, we employ a deep learning approach to discover a compromised computer account. We postulate that using user behavior knowledge to recognize the pattern of user tasks with significant features can identify user behavior. Our work provides proof of concept detector to identify the internal and external attacks in a windows operating system. To identify the objective of user operations so as to observe the user behavior by mining and summarizing the raw unorganized data. We capture our dataset from low level prepared and cleaned in a significant approach. We seek to identify significant physical features of valid users so as to oppose abnormal attack by using deep learning.

Published

2020-12-01

Issue

Section

Articles