Skripsi
ANALISIS PERBANDINGAN ALGORITMA SUPPORT VECTOR MACHINE DAN EXTREME GRADIENT BOOSTING UNTUK DETEKSI INTRUSI JARINGAN PADA DATASET NSL-KDD DAN UNSW-NB15
As cyber threats become increasingly complex, machine learning-based Intrusion Detection Systems (IDS) are crucial for network security. This study compares the performance of Support Vector Machine (SVM) and Extreme Gradient Boosting (XGBoost) using a classic dataset (NSL-KDD) and a modern dataset (UNSW-NB15). The methodology encompasses exploratory data analysis, data preprocessing, baseline model testing, and hyperparameter tuning. The evaluation results on NSL-KDD show comparable performance between both algorithms with an accuracy of 81%, though heavily dominated by False Negative errors (failed detections). Conversely, on the UNSW-NB15 dataset, XGBoost demonstrates superior and consistent performance with an accuracy of 86%, whereas SVM experiences a drop in precision due to an increased dominance of False Positives (false alarms). In terms of computational performance, XGBoost proves its scalability by completing training within seconds, far outperforming SVM, which requires up to 54 minutes. This study concludes that XGBoost is more accurate, efficient, and robust than SVM for large-scale network security dynamics.
| Title | Edition | Language |
|---|---|---|
| PENERAPAN MACHINE LEARNING DENGAN TENSORFLOW UNTUK MENDETEKSI BAHASA ISYARAT BISINDO BERBASIS APLIKASI ANDROID | id |