Classification model for accuracy and intrusion detection using machine learning approach.

Area of Science:

• Cybersecurity
• Machine Learning
• Network Security

Background:

• Increasing internet demand heightens network security concerns.
• Intrusion Detection Systems (IDS) are crucial for mitigating diverse network attacks like DDoS, ransomware, and botnets.
• Effective IDS require algorithms capable of accurately detecting and predicting malicious network activities.

Purpose of the Study:

• To evaluate and compare the performance of Naïve Bayes (NB), Support Vector Machine (SVM), and K-nearest neighbor (KNN) algorithms for network intrusion detection.
• To identify the most suitable machine learning algorithm for enhancing IDS accuracy and reducing processing time.
• To utilize performance metrics for selecting the best-fit algorithm for future intrusion behavior prediction.

Main Methods:

• Applied three distinct classification machine learning algorithms: Naïve Bayes (NB), Support Vector Machine (SVM), and K-nearest neighbor (KNN).
• Utilized the UNSW-NB15 dataset for training and testing the algorithms.
• Generated classification reports (Precision, Recall, F1-score) and confusion matrices to assess performance.

Main Results:

• The K-nearest neighbor (KNN) algorithm exhibited superior accuracy and efficiency in detecting network intrusions compared to Naïve Bayes and Support Vector Machine.
• Performance metrics indicated KNN's effectiveness in learning patterns of suspicious network activities.
• Comparative analysis of classification reports and confusion matrices validated the chosen algorithm's performance.

Conclusions:

• The K-nearest neighbor (KNN) algorithm is the most effective among the tested algorithms for enhancing network Intrusion Detection Systems.
• The study provides a validated approach for selecting and implementing machine learning models in IDS for improved network security.
• Findings support the use of KNN for real-time prediction and analysis of future intrusion behaviors.