An Intrusion Detection Framework Using an Ensemble of Deep Belief Networks and Random Forest Classifiers
Tasfia Anwar, Mohammed Moshiul Hoque · 2024
Intrusion detection is an essential aspect of information security as the danger of network intrusion has increased due to the quick expansion in amount and complexity of network traffic data. Accurate identification of various network threats is essential, but traditional intrusion detection methods need help to process data from fast networks and often fail to recognize the newly emerging attacks. This article suggests a unique method for intrusion detection that combines Random Forest (RF) and Deep Belief Networks (DBN) in order to address these issues. This innovative ensemble technique leverages the strengths of DBN (a deep model) and RF (a shallow model) to create a powerful solution. Evaluation results demonstrate the effectiveness of this approach, with the ensemble achieving the highest weighted F1-score (0.96) among all tested models, underscoring its potential to enhance intrusion detection capabilities.