Enhanced Bayesian network classifier for intrusion detection: integrating semilazy learning and multiconditional entropy

In the rapidly evolving landscape of cybersecurity, the need for sophisticated intrusion detection is paramount. Traditional methods often struggle to keep pace with the complexity and diversity of emerging threats. This paper presents an enhanced Bayesian network classifier (EBNC), integrating a semi-lazy learning framework and multi-conditional entropy to enhance the accuracy and efficiency of intrusion detection. The EBNC dynamically constructs class-specific local classifiers for each testing instance, thereby optimizing the decision-making process. It also employs multi-conditional entropy to quantify causal and conditional dependencies implicated in each testing instance. Experiments conducted on the NSL-KDD dataset demonstrate that the EBNC outperforms alternative algorithms with respect to the 0-1 loss, bias, F1 score, as well as training and classification time. These results highlight the EBNC's efficiency and accuracy in intrusion detection, demonstrating its adaptability to new threats and its capacity for rapid, informed responses in the expanding field of cybersecurity.