Related Experiment Video
Updated: Jan 17, 2026
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
1.0K
A dual-phase deep learning framework for advanced phishing detection using the novel OptSHQCNN approach
Srikanth Meda1, Vangipuram Sesha Srinivas2, Killi Chandra Bhushana Rao3
1Department of Computer Science and Engineering, RVR&JC College of Engineering, Guntur, Andhra Pradesh, India.
Peerj. Computer Science
|September 24, 2025
Summary
This study introduces a new OptSHQCNN method for detecting phishing websites, achieving over 99% accuracy. This deep learning approach enhances online security by accurately identifying malicious sites.
Area of Science:
- Cybersecurity
- Machine Learning
- Deep Learning
Background:
- Phishing attacks are a prevalent cyber threat, compromising network security and leading to financial loss and data theft.
- Existing anti-phishing tactics are often imprecise and ineffective against sophisticated phishing attempts.
- Deep Learning (DL) offers a promising approach for accurately identifying intrinsic website features to detect phishing.
Purpose of the Study:
- To propose a novel phishing detection method, OptSHQCNN, leveraging deep learning and optimization algorithms.
- To enhance the accuracy and effectiveness of phishing detection systems.
- To develop a robust solution for identifying and mitigating phishing threats in real-time.
Main Methods:
- A two-phase methodology: pre-deployment (data preprocessing, feature extraction using Convolutional Block Attention Module (CBAM), feature selection via Red Kite Optimization Algorithm (RKOA)) and post-deployment (URL encoding with Optimized Bidirectional Encoder Representations from Transformers (OptBERT)).
- Classification using a Shallow hybrid quantum-classical convolutional neural network (SHQCNN) model.
- Hyperparameter tuning of the SHQCNN model using the Shuffled Shepherd Optimization Algorithm (SSOA).
Main Results:
- The proposed OptSHQCNN method achieved exceptional performance, with accuracy, precision, recall, and F1-score exceeding 99%.
- The technique demonstrated superior performance compared to existing popular phishing detection methods.
- The findings support the development of security plugins for various applications and devices.
Conclusions:
- The OptSHQCNN method presents a highly accurate and effective solution for phishing detection.
- The study highlights the potential of integrating advanced DL models and optimization algorithms for cybersecurity.
- The developed approach can significantly contribute to enhancing user security against evolving phishing threats.
