Ddos Attacks Recognition And Forecasting Using Machine Learning Algorithms

• Author(s):

  Vooradi Sandya | B. Kinshu | D.Vithin Sai | M. Nanda Varma

• Keywords:

  Machine Learning, DDoS Attacks, Research, XGBoost.

• Abstract:

  Distributed Network Disruptions, Often Termed Distributed Denial Of Service (DDoS) Attacks, Exploit Vulnerabilities In System Infrastructures, Such As Web Servers Of Legitimate Organizations. Traditional Research Has Predominantly Utilized Outdated Datasets Like The KDD Dataset, Which May No Longer Represent The Evolving Nature Of Modern Threats. To Address This Gap, The Present Study Employs A Contemporary Dataset And Machine Learning Techniques To Identify And Forecast Various Forms Of DDoS Attacks.This Research Implements A Comprehensive Detection Framework Using Two Advanced Classification Algorithms: Random Forest And XGBoost. The UNSW-NB15 Dataset, Obtained From A GitHub Repository, Served As The Foundation For Training And Testing, While Python Was Used For Simulation And Model Deployment.Performance Evaluation Was Conducted Using A Confusion Matrix. In The First Experiment, The Random Forest Model Achieved A Precision And Recall Of 89%, With An Overall Accuracy Of 89%, Indicating Robust Performance. In The Second Test, The XGBoost Classifier Slightly Outperformed, Attaining Approximately 90% In Both Precision And Recall, With An Accuracy Of 90%.Compared To Previous Studies, Which Reported Lower Accuracy Rates Of 85% And 79%, The Proposed Methodology Demonstrates A Significant Enhancement In Identifying And Classifying DDoS Attacks, Offering A More Effective And Modern Solution To A Growing Cybersecurity Concern.

---

Other Details

• Paper id:

  IJSARTV11I6103840

• Published in:

  Volume: 11 Issue: 6 June 2025

• Publication Date:

  2025-06-30

Download Article