Multi-Vector Attack (MVA) is the utilization of various attack techniques and methods on a single target. The inability of existing traditional methods in mitigating this attack is a major problem, this poses a huge threat to individuals and organizations. This study improves the detection of MVA using a packet capture (PCAP) dataset through Multilayer Perceptron (MLP) Network. The proposed solution was implemented using python 3 programming language running on google Collaboratory GPU. This study uses a dataset containing 1,047,908 PCAP instances used in training the models. To evaluate the proposed solution, a comparative analysis of the proposed solution and three machine learning models were done based on training time, detection accuracy and F1-score. Despite the fact that the MLP model was train with input neuron, hidden neuron and output neuron of 150, 100 and 50 respectively over 100 epochs, the MPL classifiers was competitive to gradient boost, KNN and random forest machine learning algorithm in terms of detection accuracy. However, the MPL classifier struggles compared to the three machine learning algorithm in terms of training time. The evaluation result of the proposed solutions reports 99.85% detection accuracy, F1-Score of 99% which was verified using the multiclass confusion matrix.
Machine Learning, Anomaly Detection, Multi-Vector Attacks, Information Security, Intrusion Detection, Multilayer Perceptron.
IRE Journals:
Ipole-Adelaiye Nancy , Fori Barka Tatama , Onu Egena , Maikori Jenom , Lawal Ibrahim
"Detecting Multi-Vector Attack Threats Using Multilayer Perceptron Network" Iconic Research And Engineering Journals Volume 8 Issue 1 2024 Page 119-123
IEEE:
Ipole-Adelaiye Nancy , Fori Barka Tatama , Onu Egena , Maikori Jenom , Lawal Ibrahim
"Detecting Multi-Vector Attack Threats Using Multilayer Perceptron Network" Iconic Research And Engineering Journals, 8(1)