Cybersecurity in Network Traffic: Integrating Statistical Techniques with AI

Abstract

This study investigates a hybrid model combining statistical components with AI models to support cybersecurity against attacks on network traffic. To overcome the restrictive nature of classical models such as signature based intrusion detection, this research integrates statistical methods like Z-scores and stack counting with AI techniques including logistic regression and neural network. The hybrid approach is recognized for its ability to identify anomalies, classify network traffic, and respond to changing cyber threats. Results from testing on synthesized data showed that the model was able to detect anomalies to a large extent, as there was considerable accuracy, precision, and recall in targeting malignant traffic. This unified pipeline connects traditional data analytical methods and machine learning in a manner that can expand to cater for real time cyber security needs.