A Review Study on Advanced Hybrid System for URL Threat Detection

With the exponential growth of web-based services, the security of online resources has become increasingly critical. Malicious URLs pose a significant threat to both individuals and organizations, often leading to data breaches, malware infections, and phishing attacks. Traditional detection methods struggle to keep up with the evolving techniques used by cybercriminals. This paper presents a comprehensive review of advanced hybrid systems that combine multiple machine learning models for effective URL threat detection. Specifically, we focus on hybrid systems incorporating Decision Trees (DT), Support Vector Machines (SVM), and Naive Bayes (NB) models, which have shown promising results in handling the complexities of URL classification. Decision Trees offer interpretability and flexibility, while SVM excels in highdimensional feature spaces, and Naive Bayes provides simplicity with strong probabilistic foundations. By combining these models, hybrid systems can leverage the strengths of each to achieve higher accuracy, faster detection, and improved generalization. The paper reviews various hybrid approaches, evaluates their performance, and identifies challenges and opportunities for future research in URL threat detection. Our study aims to provide a thorough understanding of the synergies between these models and their potential for building more robust, scalable, and real-time security systems for web applications.