<![CDATA[This study aims to develop a machine learning-based cyber attack anomaly detection system on the Academic Information System (SIAKAD) server logs of STKIP DDI Pinrang using the Random Forest algorithm. The research employed a quantitative experimental approach using server log data from July 2022 to December 2024, complemented by brute force and HTTP flood attack simulations conducted from January to March 2025. The research stages included log preprocessing, behavioral feature extraction, data labeling, model training, and system performance evaluation. The main features used were request rate, error ratio, path diversity, and access frequency to the sensitive /lupapassword endpoint. The results showed that the Random Forest algorithm achieved an accuracy of 98.3%, precision of 97.5%, recall of 96.8%, and F1-Score of 97.1% with a False Positive Rate of 1.2%. In addition, the system demonstrated an average inference time of 28 ms, enabling near real-time detection without affecting server performance. This study proves that Random Forest is effective as a cyber attack anomaly detection solution for academic environments with limited technological infrastructure. ]]>
Copyrights © 2026
