Machine Learning-Based Malicious Web Page Detection
Machine Learning-Based Malicious Web Page Detection

Summary/Abstract: With the growing spread of the internet, malicious activities like phishing attacks have become a major cybersecurity threat. To fill the gap in the current research on identifying such attacks, which aim to steal personal and financial data by redirecting users to fake online sites, this study sought to develop a Machine Learning (ML) model that can detect malicious websites with high accuracy using URL information. For this, Convolutional Neural Network (CNN) architecture was used that analyzes URL strings with high precision. Training and testing the model resulted in a validation accuracy of 89.95%. The developed CNN-based model has the potential to be integrated into cybersecurity systems to proactively defend users against phishing, as this experimental result clearly shows. Hence, this study makes an important contribution to improving digital security by demonstrating the effectiveness of CNN in detecting malicious websites based on URLs.

• Details
• Contents