Machine Learning Approaches for Phishing Detection Using URL Analysis
Diyana Kinaneva, Georgi Hristov, Georgi Georgiev, Plamen Zahariev · 2025
In recent years, the Internet has become an essential part of our daily lives. Globally, the digital population, which includes Internet and social media users, numbers 5.44 billion, with over 90% using social media. Various events of the past decade, such as COVID-19, have accelerated the consumption of digital services, increasing the need for digital education, commerce, and work. However, the security of data exposed in the public domain remains a critical concern. Network security, a concept, strategy, and mechanism, has existed as long as networks have existed. As long as information is shared, attempts at fraud, theft, and attacks will persist. This publication examines various approaches to mitigating the misuse of personal data and mechanisms to protect against such threats. Multiple datasets have been consolidated, and a Random Forest classifier has been applied on the newly constructed dataset. Although the accuracy of the related models is high enough, the work has been continued with the efforts to improve the dataset and to polish the accuracy of the model. With the newly proposed dataset the accuracy reached 98%. Artificial intelligence plays a crucial role in enhancing cybersecurity measures, but it also facilitates cyberattacks. This dual nature suggests the use of artificial intelligence for both defensive and offensive purposes in the cyber domain.