<![CDATA[Phishing attacks are a cybersecurity threat often used to steal sensitive user information through fake websites that resemble legitimate sites. Therefore, this study aims to analyze and compare the performance of the Random Forest and Convolutional Neural Network (CNN) algorithms in detecting phishing websites based on Uniform Resource Locator (URL) features. The dataset used was obtained from Web Application Firewall (WAF) security logs on the network infrastructure at XYZ University, which record URL access activities on the web system. The data was then processed and labeled into two categories: phishing and legitimate websites. The dataset used in this study consists of 549,346 URL records. The research stages include data exploration (Exploratory Data Analysis / EDA), URL text preprocessing, character-based feature extraction and URL tokenization, and model training using the Random Forest algorithm and a 1D Convolutional Neural Network (1D-CNN) architecture. Model evaluation was conducted using accuracy, precision, recall, and F1-score metrics as well as confusion matrix analysis. The results showed that the Random Forest model achieved an accuracy of 82.69%, while the 1D-CNN model achieved a higher accuracy of 95.94%. Furthermore, the CNN training process demonstrated a steady increase in accuracy and a decrease in loss values in each epoch. Based on these results, it can be concluded that the deep learning approach using CNN outperforms the Random Forest method in detecting URL-based phishing websites]]>
Copyrights © 2026
