Scholarlink

Conventional intrusion detection methods struggle to address the growing complexity of cyber threats. To enhance the detection of cyberattacks, this study employs machine learning techniques, specifically Random Forest, Decision Tree, Support Vector Machine, Naive Bayes, and a proposed Hybrid model. It evaluates performance through various metrics using the CSE-CIC-IDS2018 dataset, demonstrating that Hybrid and Tree-based models perform well in terms of accuracy and robustness. In this research, it has been found that standardization plays a vital role in optimizing SVM output results. The findings highlight the importance of feature selection, ensemble learning, and data preprocessing in developing scalable and adaptable cyber threat detection solutions.

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

Machine Learning for Cyber Attack Detection: Insights Into Model Performance and Optimization

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Caressa Wandaphi Pyngrope,
  • Tanmoy Das,
  • Maverick Yenkokpam,
  • Cherukuri Venkta Sai Narendra,
  • Bobby Sharma

摘要

Conventional intrusion detection methods struggle to address the growing complexity of cyber threats. To enhance the detection of cyberattacks, this study employs machine learning techniques, specifically Random Forest, Decision Tree, Support Vector Machine, Naive Bayes, and a proposed Hybrid model. It evaluates performance through various metrics using the CSE-CIC-IDS2018 dataset, demonstrating that Hybrid and Tree-based models perform well in terms of accuracy and robustness. In this research, it has been found that standardization plays a vital role in optimizing SVM output results. The findings highlight the importance of feature selection, ensemble learning, and data preprocessing in developing scalable and adaptable cyber threat detection solutions.