TITLE:
Effectiveness of Deep Learning Algorithms in Phishing Attack Detection for Cybersecurity Frameworks
AUTHORS:
Mitra Penmetsa, Jayakeshav Reddy Bhumireddy, Rajiv Chalasani, Srikanth Reddy Vangala, Ram Mohan Polam, Bhavana Kamarthapu
KEYWORDS:
Cybersecurity, Phishing Attacks, Machine Learning, Deep Learning (DL), GRU, PhishTank Data, Cyber Attack Defense
JOURNAL NAME:
Journal of Data Analysis and Information Processing,
Vol.13 No.3,
August
28,
2025
ABSTRACT: The widespread use of internet technologies is limited because people are worried about cybersecurity. With phishing, cyber criminals pose as reputable entities to trick users and access important information. Standard detection approaches are difficult to follow along with the constantly changing strategies of cybercriminals. A new phishing attack detection framework is presented in this research, using the Gated Recurrent Unit (GRU) Artificial Intelligence (AI) model. Labels have been added to the Uniform Resource Locators (URLs) in the PhishTank dataset, so the model learns what is phishing and what is not. A good data preprocessing method involving feature extraction, dealing with missing data, and running outlier detection checks is applied to maintain high data quality. The performance of the GRU model is outstanding, reaching 98.01% accuracy, F1-score of 98.14%, 98.41% recall, as well as 98.67% precision, better than that of classical Machine Learning (ML) methods, including Adaptive Boosting (AdaBoost) and Long Short-Term Memory (LSTM). The proposed approach correctly handles dependencies among elements in a URL, resulting in a strong method for detecting phishing pages. Results from experiments verify the model’s potential in accurately identifying phishing attacks, offering significant advancements in cybersecurity defense systems.