Enhancing Cybersecurity Through Machine Learning: Differentiating Browsing Patterns to Identify Potential Threats
Charulatha Tammana, Sriya Komaragiri, Gayathri Munnella, Hemalatha Jalapati, Sravan Kumar Gunturi · 2024
Browsing history is a crucial element of an individual’s digital footprint, meticulously documenting online activities such as website visits and search queries. This wealth of data not only reveals personal interests and behaviors but also carries significant implications for privacy and cybersecurity. A developed machine learning model utilizes this data to distinguish between the browsing patterns of legitimate users and potential attackers. This innovative approach has the potential to revolutionize organizational security efforts by enabling proactive identification of suspicious activities. In a controlled testing environment, the model demonstrated impressive performance, achieving an accuracy of 97.2%, with the Extra Trees classifier exhibiting the highest precision at 98.5% and a recall rate of 97.0%. These results underscore the model’s capability to effectively pinpoint potential threats, thereby enhancing cybersecurity measures and reducing the risk of breaches. By analyzing deviations from normal browsing patterns, this technique represents a valuable addition to cybersecurity strategies, aimed at protecting critical infrastructures against increasingly sophisticated threats. This approach, therefore, offers significant potential for improving security protocols by providing a more targeted and efficient means of threat detection and prevention.