Securing IoT with Advanced Machine Learning and deep leaning Based Intrusion Detection Systems

Main Article Content

Mohit Jain
Dr Saurabh Mandloi

Abstract

Intrusion detection systems (IDS) are essential for protecting networks from cyberattacks in the connected digital world of today. Using the well-known KDD CUP 99 dataset, this paper investigates the usage of Support Vector Machine (SVM), Random Forest methods, and VGG16 AND VGG19 for intrusion detection. The dataset, which includes a variety of normal and attack traffic patterns, offers a thorough benchmark for assessing machine learning-based IDS approaches. While Random Forest is used for its ensemble learning technique, with vgg16 and vgg19 increasing accuracy, the SVM model is used for its capacity to identify the best hyper planes for classification. Because of its interpretability and simplicity, the machine learning and deep learning models are also used to categorize network traffic. To improve model performance, data pre-processing techniques include categorical encoding, dimensionality reduction, and feature normalization. The outcomes of these tests show that while both models attain high detection accuracy, vgg19 performs better than the others in terms of precision and recall because of its enhanced capacity to manage intricate datasets and minimize overfitting. The results highlight how deep learning methods, especially vgg19, have the potential to improve the identification of known and new attack pathways.

Article Details

Section
Articles