International Research Journal of Education and Technology

The project introduces a web-based hybrid detection system that has been created to enhance cyber security against contemporary threats like phishing, malware, ransomware, and Trojans. Conventional security mechanisms tend to fail when confronted with advanced cyberattacks, mainly because they rely on static rules. The system overcomes that drawback by combining both supervised and unsupervised machine learning algorithms. Random Forest and XGBoost are utilized for precise URL classification and identification of malicious activity, and K-Means and Isolation Forest are employed for powerful anomaly detection. Scalability and simplicity are built into the system, with a user-friendly dashboard appropriate for both technical and non-technical users. Real-time notifications and detailed log reports facilitate effective threat monitoring and response. Through its integration of both conventional and state-of-the-art detection techniques, the system offers an effective proactive cybersecurity solution that learns on its own, evolves to face new threats, mitigates false positives, and increases organizational resilience in an ever-changing digital world.

Key Words: Hybrid detection, Random Forest, XGBoost, K-Means, Isolation Forest, URL classification, Trojan detection, anomaly detection, cybersecurity.