In the modern digital world, cybersecurity threats are increasing rapidly, making the protection of network infrastructures critically important. Traditional rule-based intrusion detection systems often struggle to adapt to evolving attack patterns and sophisticated network intrusions. This project proposes an Al-driven, adaptive network attack detection system that leverages machine learning techniques to analyze NetFlow data and accurately identify potential security threats. NetFlow, which captures metadata about network traffic flows, provides valuable information like source and destination IPs, ports, packet counts, and flow durations. By extracting and analyzing these features, the system can detect abnormal behaviors and differentiate between normal and malicious activities.  This approach not only enhances detection accuracy but also adapts dynamically predicting attacks. The system aims to strengthen network defenses, reduce the risk of breaches, and improve the overall security posture of organizations by providing intelligent intrusion detection capabilities.

Keywords- cyber security, machine learning, anomaly detection, predictive analytics, network traffic analysis, cyber threats.