Machine Learning for Cloud-Based DDoS Attack Detection: A Comprehensive Algorithmic Evaluation

Distributed denial of service attacks, in the recent times, have become a critical threat to not only private server connections but also meta scale cloud computing environments. This research paper furnishes a detailed analysis of five machine learning algorithms in context with network intrusion detection that are; Isolation Forest, Random Forest, Support Vector Machine (SVM) and Artificial Neural Networks (ANN). The dataset used in this study, to differentiate between attack and normal cases, is the KDD cup 1999, wherein the dataset has been converted into binary classes to differentiate between the attack and normal cases. Following are the part of the research; Preprocessing the data, Scaling the features, Training the model and evaluating the performance using crucial metrics like accuracy, precision and F1 score and recall. Graphical representation and confusion matrices for data visualization have been made use of in addition to the aforementioned. The results show a penetrative information about the relative effectiveness and success of different ML algorithms in identifying network intrusions and hence can be used as insightful references for academics and practitioners in the field of network security. A well- explained algorithm selection, based on performance indicators enables an advanced approach to the study, thus providing a framework for further developments in systems of intrusion detection.