Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 58087 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Jalili, Rasool
- Abstract:
- Insider threats are one of the most challenging threats in organizations, as individuals can escape the firewalls, intrusion detection systems, and other security mechanisms. People with managerial salaries can also perform devastating behavior against the organization using existing permits. Submission to computer systems, theft of valuable information, and bypassing organizational procedures are insider threats. Therefore, the need to pay attention to these threats is of particular importance. In this thesis, the aim is identifying insider threats on sensitive data. Sensitive data are either identified by the database manager or are known at the time of production due to sensitivity. Employee activity logs in the organization are one of the critical data set of each organization. Therefore, in this thesis, the identification of insider threats is based on raw logs of the organization. The proposed method for detecting insider threats includes various stages of monitoring, data analysis, processing of employee raw logs, and the use of unsupervised machine learning algorithms. The results of the proposed method show that an autoencoder is the best combination of anomaly detection using databased representations. Temporal data representations in the form of percentile makes significant improvements over the main extracted data. Therefore, it makes it possible to detect the threats under very low investigation budget and is well generalized to the new data. It also works better than the previous work done, and is capable of generalizing work in different environments
- Keywords:
- Anomaly Detection ; Sensitive Data ; Deep Learning ; Security Threats ; Insider Threats ; Intrusion Detecticn and Prevention
Digital Object List
-
محتواي کتاب
- view
Bookmark