Several security issues aligned with web application are mostly relative to intrusions. These latter are due to the trend and fast development of Web Applications (Web-App). To reduce the risks associated with Web-App problems, the developers must take into consideration numerous measures by elaborating highly secured applications in order to avoid the known attacks. In case such measures fail to serve their purpose, it is imperative to detect and identify the source of these attacks to minimize the estimated risks. The Intrusion detection is one of most effective techniques that enables developers to identify and prevent system damage. Most of defensive techniques of Web intrusion detection are not capable of recognizing and managing the complexity of Web-App attacks. Nevertheless, the principle of Machine Learning (ML) approach allows efficiently to detect the known and the unknown attacks. In this thesis, we are studying and discussing ML techniques to classify HTTP requests in the dataset known as CSIC 2010 HTTP of which consists of normal or abnormal traffic. These experiences generate results for several ML techniques that overlaps and different set of functionalities, which allow to compare the quality of different sets of functionalities between different ML techniques, at least, under the dataset mentioned previously.
| Date | 28 Oct 2021 |
|---|
| Original language | French |
|---|
| Awarding Institution | - École de technologie supérieure
|
|---|
| Supervisor | Chamseddine Talhi (Supervisor) |
|---|
Lagrini, O. (Author),
Talhi (Supervisor),
28 Oct 2021Student thesis: Master's thesis › Master in Engineering: Information Technology Engineering