In recent years the development of the internet and its applications has increased rapidly, and its use occupies continuously more and more a larger part of people daily lives. Today, the internet is a basic and necessary means for communication, entertainment, information, shopping and many other functions that are now done through it. Unfortunately, with the development of these features, illegal activities as cheating other users, accessing confidential and secret information, promoting certain products and even interrupting the deployment of websites from the internet, have also increased, since hackers are exploiting the vulnerabilities in the security of web applications and systems. Cybersecurity focuses on the development of protection systems and methods that aim to detect and identify an impending cyber-attack, thus contributing drastically to protection against malicious actions. On the other hand, the field of Machine Learning focuses on developing techniques that allow a computing system to "think" and "decide", and not just explicitly execute commands that have been dictated to it by the programmer. The field of Machine Learning is widely used in various domains, such as cybersecurity, which this dissertation deals with. In the context of this dissertation, a system was modeled and developed to receive necessary and useful information about user behavior in an online e-commerce application, and after storing and processing the data in a specific way, finally feeding them into a sequential classification machine learning model to characterize the user behavior as either benign or malicious.