Scholarlink

This paper proposes a novel Intrusion Detection and Prevention System (IDPS) that employs machine learning techniques to bolster network security. By leveraging labelled datasets such as CIC-IDS2017 and CIC-IDS IOT 2023, the system undergoes rigorous data preprocessing to extract meaningful features. A comprehensive ensemble of supervised learning models, including Random Forest, XGBoost, CNN, LSTM, KNN, and Model Stacking, is trained and evaluated for intrusion detection accuracy. Additionally, unsupervised clustering algorithms (K-Means, DBSCAN) are integrated to identify anomalous network traffic patterns. Experimental results demonstrate the efficacy of the proposed IDPS in detecting and preventing cyber threats, particularly within the evolving 5G ecosystem.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Real-Time Network Intrusion Detection System Using Machine Learning

  • 章节

  • 发表日期：

    2025年

查看全文
  • Sam Peter,
  • J. L. Aravind,
  • Feba Mariyam Jacob,
  • Johann Varghese George,
  • Tessy Mathew

摘要

This paper proposes a novel Intrusion Detection and Prevention System (IDPS) that employs machine learning techniques to bolster network security. By leveraging labelled datasets such as CIC-IDS2017 and CIC-IDS IOT 2023, the system undergoes rigorous data preprocessing to extract meaningful features. A comprehensive ensemble of supervised learning models, including Random Forest, XGBoost, CNN, LSTM, KNN, and Model Stacking, is trained and evaluated for intrusion detection accuracy. Additionally, unsupervised clustering algorithms (K-Means, DBSCAN) are integrated to identify anomalous network traffic patterns. Experimental results demonstrate the efficacy of the proposed IDPS in detecting and preventing cyber threats, particularly within the evolving 5G ecosystem.