Penetration testing procedure using machine learning

Abstract

The main aim of this study is to determine the effectiveness of a penetration testing tool, GyoiThon as a Machine Learning tool by conducting penetration tests on websites, including the websites that have Content Management System (CMS) frameworks, to identify their vulnerabilities and assess the effectiveness of GyoiThon features in penetration testing. This experiment will determine how well the automation framework is executed for penetration testing. This research hypothesized, that if the feature of a penetration tool consists of any form of Machine Learning algorithm, the more effective the feature can search for more vulnerabilities. To achieve the aim of this paper, an experiment was conducted to evaluate the effectiveness of the two features of GyoiThon, the Default and Machine Learning modes. In the end, it was revealed that the Machine Learning mode of GyoiThon discovered more types of vulnerabilities than using the Default mode of GyoiThon, proving the hypothesis to be right.