Security situation awareness algorithm of network information transmission based on big data.

Area of Science:

• Cybersecurity
• Machine Learning
• Big Data Analytics

Background:

• Traditional security mechanisms are challenged by increasingly complex and diverse malicious network attacks.
• Effective threat identification in network information transmission is crucial for maintaining network security.

Purpose of the Study:

• To propose a new three-stage algorithm for network security situation awareness.
• To enhance the identification of threats in network traffic using machine learning and big data processing.

Main Methods:

• Feature vector creation from network traffic flows (statistical, temporal, flow-based, relational features).
• Hybrid feature selection using Distributed K-Means (D-KMeans) for clustering and Mutual Information (MI) analysis.
• Network traffic classification using a Distributed K-Nearest Neighbor (D-KNN) model.

Main Results:

• The algorithm achieved 98.91% accuracy, 93.71% precision, 98.95% recall, and 96.00% F-Measure on the CICIDS2017 dataset.
• Demonstrated a statistically significant increase of at least 1.2% in accuracy compared to state-of-the-art approaches.
• Validated the effectiveness and efficiency in identifying security threats in large-scale network environments.

Conclusions:

• The proposed three-stage algorithm offers a robust and scalable solution for network traffic classification and threat identification.
• The integration of machine learning and big data processing significantly improves network security awareness.
• This approach provides an effective means to combat sophisticated malicious network attacks.