Development of machine learning based attack detection system using web logs

With the rapid development of technology, communication and interaction between people has gained great momentum. As distances are shortened, many transactions such as shopping are moved to a different size and started to change shape. Social life has even been carried out in virtual environments that are online. More and more people are using more information systems. In usage of any type of tool such as information systems, always there have been people whom trying to disturb the system or to use it for a different purpose as well as well-intentioned people. For this reason, the widespread use of computer systems has made it necessary to record the transactions of people using these systems without entering personal information. Nowadays, most of the internet web pages are prepared for different purposes and prepared for people. All of the server systems, whether restricted access or publicly accessible pages, keep users' access log. These log files contain non-personal access information. In this thesis, an attack detection application was developed for the related system via the log files stored in the server systems. The application has been developed on Inonu University web site logs.