|
|
|
AUTHOR(S): Dr Babangida Zubairu, Aisha Ibrahim Gide
|
|
TITLE |
 PDF |
|
ABSTRACT A Mobile Ad-hoc Network (MANET) is a dispersed, decentralized network of mobile wireless nodes that interacts directly with one another without the use of centralized management or stationary base stations. In a MANET's, nodes are always moving, randomly and unpredictably, which presents a number of difficulties and leaves these networks especially open to different security risks. These infrastructure-less networks are especially vulnerable to security threats like Denial of Service (DoS) attacks, black hole attacks, network partitioning, and node selfishness since they lack central management and have limited hardware resources. This paper presents a framework for detecting and mitigating Denial of Service (DoS) attacks in Mobile Ad-hoc Networks (MANETs) using various machine learning models, including Support Vector Machine (SVM), Random Forest, Neural Network, and K-Nearest Neighbors (KNN). In order to simulator both normal and attack traffic, the proposed system generate synthetic datasets from simulation, trains these models, and assesses their efficacy using a number of performance metrics, such as accuracy, scalability, energy consumption, resource utilization, and network throughput. The results shows that Random Forest, SVM, KNN, and Neural Networks have the highest detection accuracy. Neural Networks also showed better network performance and scalability, which made them appropriate for high-traffic situations. On the other hand, KNN is observed to have less energy consumption, which made it an effective choice for situations where resources are limited. |
|
KEYWORDS DoS Attack Detection, Mobile Ad-hoc Networks (MANETs), Machine Learning Models, Neural Network |
|
|
|
Cite this paper Dr Babangida Zubairu, Aisha Ibrahim Gide. (2025) Machine Learning-Based Framework for Detecting and Mitigating DoS Attacks in Mobile Ad-Hoc Networks (MANETs). International Journal of Computers, 10, 151-160 |
|
|
