Advancements in Intrusion Detection Systems: A Comparative Analysis of Machine Learning and AI Methodologies

To improve the effectiveness of intrusion detection system (IDS), this article explores the use of machine learning and AI based technology. The study encompasses the application of various approaches in the category of classification, clustering, and neural networks, across diverse datasets. The pr...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:International Conference on Computing, Communication, and Networking Technologies (Online) s. 1 - 7
Hlavní autoři: Panigrahi, Swayankar, Mohanty, Soumya Ranjan, Dalai, Ritik Roshan, Wagh, Mamata P., Sawaika, Anuradha, Pradhan, Priti Priyadarsani
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 24.06.2024
Témata:
Accuracy
Classification Algorithms
Clustering Algorithms
Computational modeling
Computer hacking
Convolution
Cyber Security
Deep Learning
Feature Selection
IDS
Intrusion detection
Iterative methods
Machine learning
Malware
Nearest neighbor methods
Neural networks
ISSN:2473-7674
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:To improve the effectiveness of intrusion detection system (IDS), this article explores the use of machine learning and AI based technology. The study encompasses the application of various approaches in the category of classification, clustering, and neural networks, across diverse datasets. The primary focus is on evaluating the accuracy and efficiency of each method for consolidated analysis and detailed insightful comparison. The outcomes of this detailed study are then compared and scrutinized to extract insights into the merits and demerits of each technique. After analyzing the various techniques, it is observed that in the classification category K-Nearest Neighbors (KNN) is outperforming by showing the highest accuracy among all the classification techniques studied in this work. Moreover in clustering category Balanced Iterative Reducing and Clustering using Hierarchies (BRICH) clustering technique has shown highest accuracy. Further in the AI category all the models those are studied in this work has shown the accuracy of 91% and above. Among these AI models the Convolution Neural Network (CNN) has recorded the highest accuracy of 96% on an average across the various datasets. The overarching objective is to elevate the overall effectiveness of IDS in identifying and countering cyber threats. The findings from this study make a substantial contribution to current initiatives focused on strengthening network security and defending against the continuously evolving landscape of cyber threats.
ISSN:2473-7674
DOI:10.1109/ICCCNT61001.2024.10724759