Exploring Machine Learning Approaches for Robust Anomaly Detection and Responsive Security in IoT Frameworks
Year: 2024
Type: dissertation
Abstract: This paper evaluates machine learning (ML) and deep learning (DL) models for network traffic anomaly detection in IoT devices. Three models were tested: Support Vector Classifier (SVC), Convolutional Long Short Term Memory (ConvLSTM) and Extreme Gradient Boosting (XGBoost) Model. The motivation behind this project is to analyse, examine and evaluate machine learning and deep learning models and their effectiveness with respect to discovering and mitigating network anomalies in IoT frameworks. The accuracy for the XGBoost model is higher than the SVC and ConvLSTM models, with an accuracy of 99.98% as compared to 92.80% and 96.18% respectively. After each attack, the Inference system logs the result, sends email notifications when it detects an attack, and skips blocked IP addresses. Results demonstrate that XGBoost model has the ability to identify the DDoS attacks better, which would be promising for future network security applications in real environments.
Source: NORMA
Author: Shaju, Albin
Institution: National College of Ireland
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7066054876
