An Enhanced Distributed DoS Attack Classification in Digital Transactions using Hybrid Machine Learning Algorithms

In the financial industry, data privacy is crucial for safeguarding private customer data, stopping financial crimes, guaranteeing regulatory compliance, and safeguarding intellectual property. With distributed denial of service (DDoS) attack a major danger to privacy, the rise of digital transactions and internet usage has made protecting client data increasingly difficult. Effective and strong detection and prevention mechanisms are therefore very vital to handle this. Effective cyber-attack detection models can be developed using machine learning, therefore improving the capacity to safeguard private facts and support both scientific and pragmatic developments. Model training and evaluation made use of the CIC-IDS 2017 dataset, for its extensive network traffic facts collecting capability. Lasso was used for feature selection to boom model performance; oversampling become used thru SMote to resolve data imbalance. We used several machine learning techniques both with and without hyper parameter optimization. With 96% accuracy, a vote casting classifier performed the best among the models evaluated, therefore proving the dependability and efficiency of ensemble techniques in DDoS attack detection.