The College of Education for Pure Sciences at Wasit University discussed a master's thesis on the detection of distributed denial-of-service (DDoS) attacks based on botnets using machine learning algorithms in an Internet of Things (IoT) environment by the postgraduate student, Mr. Dawood Salman Kadhim.

The thesis aimed at developing an intelligent framework for detecting DDoS attacks based on botnets in an IoT environment by addressing the challenges of data imbalance, complex behavioral characteristics, and limited computing resources.

The thesis relied on integrating feature extraction and selection techniques using Principal Component Analysis (PCA) and Interoperability Information (MI), employing the SMOTE-ENN technique to improve data balance, and the XGBoost algorithm to achieve accurate and effective attack detection. 

The thesis demonstrated that the proposed model achieved a high level of accuracy and efficiency, with a balanced accuracy rate (B.A) of 98.33% and an F1-Macro index of 98.71%, with an inference speed suitable for application in terminal environments of 0.47 microseconds, which enhances the security of Internet of Things systems and their ability to confront modern cyber threats.