Scholarlink

This research introduces a new IDS framework based on machine learning for better learning of minority attack types of cyberthreats. To address significant class imbalance concerns, the method incorporates SMOTE-ENN, while classifiers such as random forest (RF) and XGBoost boost detection capabilities. As for the limitations of the traditional methods in identifying minority attack types, the present research process focuses on the application of the NSL-KDD dataset to solve this problem. Performance evaluations of the framework are significantly higher than those of competing algorithms in terms of accuracy, precision, recall, and F1 score. Accordingly, it can be noted that the proposed IDS framework could help in strengthening the form of “network security” for organizations by having the potential to improve the efficiency of detection and management of “cyberthreats.” The proposed methodology achieves more than 99% accuracy.

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

Smart Intrusion Detection Systems: A Novel Intrusion Detection Framework by Leveraging SMOTE-ENN with RF and XGBoost

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ratnesh Kumar Choudhary,
  • Sameer Tembhurney,
  • Khushi Barole,
  • Khushi Malik,
  • Rahul Raut,
  • Prachi Joshi

摘要

This research introduces a new IDS framework based on machine learning for better learning of minority attack types of cyberthreats. To address significant class imbalance concerns, the method incorporates SMOTE-ENN, while classifiers such as random forest (RF) and XGBoost boost detection capabilities. As for the limitations of the traditional methods in identifying minority attack types, the present research process focuses on the application of the NSL-KDD dataset to solve this problem. Performance evaluations of the framework are significantly higher than those of competing algorithms in terms of accuracy, precision, recall, and F1 score. Accordingly, it can be noted that the proposed IDS framework could help in strengthening the form of “network security” for organizations by having the potential to improve the efficiency of detection and management of “cyberthreats.” The proposed methodology achieves more than 99% accuracy.