Pengembangan Sistem Klasifikasi Serangan Siber Berbasis Machine Learning

The increasing use of web applications has led to a higher risk of cyberattacks, particularly targeting user authentication processes such as SQL Injection and Brute Force attacks. This study aims to develop a machine learning-based cyberattack classification system using the Random Forest algorithm and the CICIDS-2018 dataset. The system utilizes the Scapy package to capture network traffic on the face login of a web application implemented in a WSL Ubuntu environment. Testing was conducted under three scenarios: normal login activity until successfully accessing the dashboard as benign traffic, SQL Injection attacks performed on the login form, and Brute Force attacks simulated using a Python Script that repeatedly attempts login credentials. The Random Forest model classifies network traffic in real time and sends notifications via a Telegram Bot when an attack is detected. The results indicate that the system is capable of effectively detecting and classifying network traffic, thereby enhancing web application security.