Detection of Malicious Requests on Web Logs Using Data Mining Techniques

Sahin M. E., Ozdemir S.

4th International Conference on Computer Science and Engineering (UBMK), Samsun, Turkey, 11 - 15 September 2019, pp.463-468 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/ubmk.2019.8907087
  • City: Samsun
  • Country: Turkey
  • Page Numbers: pp.463-468
  • Keywords: malicious web requests, web application security, web access logs, data mining, machine learning
  • Hacettepe University Affiliated: No


The rapid development and deployment of Internet-based applications has created many problems related to web security. Although the number and the variety of the methods that have been developed in order to maintain security of web applications are increasing, every passing day new types of attacking methods that are targeted at these systems emerge. In order to provide a supplement method to rule and signature based security systems for web application security, it becomes more important to detect and classify the malicious web requests using machine learning techniques. In this work, we used various machine learning techniques to discover the patterns on web requests, and classified the malicious requests while comparing different techniques.