A Deep Learning-Based Innovative Technique for Phishing Detection in Modern Security with Uniform Resource Locators

Area of Science:

• Cybersecurity
• Machine Learning
• Deep Learning

Background:

• The increasing prevalence of cyberattacks, particularly phishing, necessitates advanced detection methods.
• Existing phishing detection (PD) systems require improvement to counter the growing number of malicious websites.

Purpose of the Study:

• To develop and evaluate a novel Convolutional Neural Network (CNN)-based model for highly accurate phishing website detection.
• To enhance cybersecurity by improving the precision of distinguishing legitimate from phishing URLs.

Main Methods:

• Utilized a deep learning approach, specifically a seven-layer CNN model, for classifying Uniform Resource Locators (URLs).
• Trained and tested the model on a dataset comprising 10,000 phishing and 10,000 legitimate URLs from PhishTank.
• Employed binary-categorical loss and the Adam optimizer for model training and evaluation.

Main Results:

• The proposed CNN model achieved a high accuracy rate of 98.77% in detecting phishing websites.
• The model demonstrated superior performance compared to previous state-of-the-art models, including KNN, NLP, RNN, and Random Forest.
• Experiment results indicated strong performance in terms of accuracy and a low false-positive rate.

Conclusions:

• The developed CNN-based deep learning model offers a significant advancement in phishing detection accuracy.
• The model's effectiveness in identifying phishing sites contributes to more robust cyber defense strategies.
• The approach's high accuracy and low false-positive rate make it a valuable tool for combating online phishing threats.