Adaptive Layered Machine Learning Approach to Detect and Mitigate Behavioral Based Intrusions in Wireless Sensor Network
S Saathvika, B L Accamma, Santhosh Kumar B J · 2024
In today’s dynamic cyber-security world, adjusting to emerging attack techniques necessitates creative ways for threat detection and mitigation. Using the KDD Train dataset, a decision tree-based model first differentiates between normal and unusual actions. Subsequently, employing Random Forest increases anomaly detection accuracy. K-means clustering groups attack types together utilizing a similarity index. A stacked ensemble method includes additional models such as gradient boosting and an extra tree classifier to improve accuracy even further. Detection results are recorded in a CSV file, where suspicious packet transfers trigger blockage, with transfer status has been documented. Performance assessment parameters such as accuracy, precision, recall, and F1 score are used to evaluate the model’s effectiveness in detecting anomalies and reducing threats. Validation applies to real-world datasets acquired by Wireshark and the KDD test dataset, affirming the framework’s effectiveness.