DETECTING DISTRIBUTED DENIAL OF SERVICES USING MACHINE LANGUAGE LEARNING TECHNIQUES

Vulnerabilities caused by cyberattacks impact negatively on the increased dependence of society on information and communication technologies (ICT) to conduct personal and business functions. An example of such an attack is the distributed denial of service (DDoS). This attack can disrupt business communication with clients and frustrate staff because of its potential to reduce connectivity and exchange of information between companies and their clients. To prevent these attacks, their modus operandi needs to be examined. Studies also must examine the latest trends of tactics used by DDoS attackers. The current paper aims to investigate several machine learning technologies for the detection of DDoS attacks. The accuracy of detection of DDoS attacks is examined using the CIC-DDoS dataset. Two techniques were used to preprocess the DDoS dataset to acquire the relevant features used to conduct the DDoS test. A total of 4 machine learning models have been used to detect DDoS. The results from the experiments show that the Random Forest machine learning model enabled DDoS detection with the highest accuracy of 99.997%, higher than Convolutional Neural Network (CNN), CatBoost, and Light GB. The novelty of the results is that they are based on empirical tests to determine the effectiveness of various machine learning models, thus improving the reliability and validity of the results and enhanced by the use of CIC-DDoS datasets associated with actual incidences of DDoS attacks, which makes the research framework easy to replicate to establish the validity of the findings.