Abstract:

There is a rising need for efficient methods for identifying harmful actions in computer networks due to the complexity and diversity of cyberattacks. In this study, a brand-new machine learning-based method for identifying network intrusions is presented. We suggest an elaborate structure with three stages: feature extraction, feature selection, and classification. The suggested framework analyses network traffic data using to identify patterns of suspect behaviour using various statistics and machine learning approaches. A real-world dataset is used in experiments to demonstrate the utility of the proposed methodology. The findings demonstrate the effectiveness of our approach in precisely identifying a variety of network attacks, including DoS, Remote to Local (R2L), User to Root (U2R), and probing assaults. Our methodology performs better than a number of state-ofthe- art intrusion detection methods in terms of precision, recall, and accuracy. Overall, this research helps to create approaches for spotting and preventing cyberattacks on computer networks that are efficient and scalable.

Description:

XGBoost, LSTM, SMOTE, NSL-KDD, Machine Learning