Details

Autor(en) / Beteiligte

• Parvathi Sitanaboina, Sri Lakshmi
• G, Madhuvani
• Sanjana Ch, Kusuma
• K, Savya
• M, Navya Sree

Titel

Detecting Phishing Websites: Leveraging RNNs and Domain-Specific Features for Enhanced Accuracy

Ist Teil von

• 2023 International Conference on Computational Intelligence, Networks and Security (ICCINS), 2023, p.01-07

Ort / Verlag

IEEE

Erscheinungsjahr

2023

Quelle

IEEE Electronic Library (IEL)

Beschreibungen/Notizen

• Phishing attacks pose a significant threat to online security, demanding effective detection techniques. Researchers have encountered various limitations in previous studies on phishing website detection are Limited Feature Set, Imbalance and Evolving Class Distribution and Lack of Generalizability. To address these limitations in phishing website detection, in this research paper we presents an enhanced approach for phishing website detection using Recurrent Neural Networks (RNNs) and domain-specific features. By leveraging the temporal nature of web page content through RNNs and integrating domain-specific indicators, the proposed methodology aims to improve accuracy and robustness. The study includes data collection, preprocessing techniques, RNN-based feature extraction, incorporation of domain-specific features, and training a classification model. Experimental evaluation on the UCI Phishing Websites Dataset demonstrates the effectiveness of the combined approach, showcasing its potential for combating phishing attacks. The results indicate that the integration of RNNs with domain-specific features enhances the detection system's performance compared to the baseline RNN-only approach and other counterparts. The findings contribute to the field of phishing detection and offer valuable insights for developing more reliable and efficient defense mechanisms against phishing attacks.