Aplikasi Deteksi Website Phishing Berbasis Web Menggunakan Random Forest dan Ekstraksi Fitur URL

Abstract

Advancements in information technology have raised growing concerns among various stakeholders. Phishing attacks have become one of the most common cyber threats, targeting users by imitating legitimate websites to obtain sensitive information. This study aims to develop a web-based application by implementing a supervised learning approach using the Random Forest algorithm to automatically classify URLs as phishing or legitimate. The dataset used consists of 11,054 URL instances with 30 URL-based features. The research process includes data preprocessing, feature extraction, data splitting, and classification model development and evaluation using four data partition scenarios. Model performance was assessed using accuracy, precision, recall, and F1-score as evaluation metrics. The results of the experiments show that the model achieved optimal performance with an 80:20 data split, obtaining an accuracy of 97%, precision of 97%, recall of 98%, and an F1-score of 97%. Furthermore, the trained model was implemented in a web-based application, allowing users to automatically detect URLs.