QBSO-FS method enhances firewall log analysis with optimized feature selection

Researchers have developed a novel feature selection method called QBSO-FS, which utilizes a combination of bee swarm optimization and Q-learning for improved firewall log analysis. This method was tested on the UCI Internet Firewall Data Set using classifiers such as random forest, k-nearest neighbor, and Bernoulli naıve Bayes. The QBSO-FS approach successfully identified a small subset of features that significantly enhanced classification accuracy and Macro-F1 scores, particularly for random forest on deduplicated data, achieving 99.90% accuracy and 94.88% Macro-F1. AI

IMPACT Optimizes feature selection for network security tasks, potentially improving efficiency and accuracy of firewall log analysis.

RANK_REASON The cluster contains a research paper detailing a new method for feature selection in machine learning applied to network security. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.NE (Neural & Evolutionary) →

• Bernoulli Naïve Bayes
• deep neural network
• k-nearest neighbors algorithm
• Macro F1
• particle swarm optimization
• QBSO-FS
• random forest
• UCI Internet Firewall Data Set

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →