Supervised Machine Learning for Scalable and Robust Cybersecurity A Framework for Anomaly Detection and Threat Mitigation

With the growing complexity of cyber threats, there is a critical need for improved solutions for real-time detection and mitigation. This paper explores the use of Naïve Bayes, K-Nearest Neighbors, Decision Trees, and Random Forest algorithms, all implemented using RapidMiner due to its user-friendly interface. Model performance and reliability were enhanced through essential preprocessing steps, including feature selection, normalization, and cross-validation. Google Colab was used for model training and optimization. This study highlights the importance of effective data preparation and algorithm selection in building scalable and robust machine learning models for cybersecurity. Among the evaluated models, Random Forest achieved the highest accuracy (99.04%), followed closely by KNN (98.84%).