Efficient Machine Learning Approaches for Intrusion Identification of DDoS Attacks in Cloud Networks
2025
Recent trends have highlighted that DDoS contributes to most of the general attacks on networks.Networks struggle to distinguish between legitimate and malicious transmissions.Many reasons, encompassing the intricate, inflexible, costly, and vendor-specific architecture of modern networking devices and protocols, make it difficult to develop, test, and apply DDoS techniques.The BoT-IoT dataset, an extensive network traffic resource, is harnessed by the study to train and test a Decision Tree (DT) classifier that distinguishes DDoS attack patterns.The proposed methodology involves strenuous preparation steps for the data, including missing data imputation, feature mapping, one-hot encoding, and normalization, in order to ensure preprocessed data of high quality and consistent.The DT model performed trailer, reporting an accuracy, precision, recall, and F1-score of 99.9% across major points of evaluation.As revealed by comparative analysis to baseline models (GD and RF), the DT model outperforms for DDoS activity detection.The final results accentuate the robustness, interpretability and effectiveness of the Decision Tree classifier to be an interesting solution for real-time intrusion detection systems in the cloud environment.