A Dynamic Strategy Selection Module For Anomaly Detection Wireless Sdns Based On Semisupervised Learning

Abstract

This paper suggests a dynamic strategy selection module for wireless Software-Defined Networks (SDNs) anomaly detection using Random Forest (RF) model learning techniques. The intricate and dynamic architecture of wireless software-defined networks (SDNs) makes anomaly identification with sparse tagged data extremely challenging. We offer a Random Forest (RF) model learning system that aims to increase detection accuracy by using both labeled and unlabeled network traffic data. This is intended to solve the problem that we have found. The dynamic strategy selection module continuously adjusts to changes in network traffic patterns and then chooses, in real-time, the best detection method based on the present situation. The Random Forest (RF) model learning model is trained and developed using TensorFlow, allowing the system to identify irregularities with limited labeled data. Compared with traditional supervised algorithms, the suggested solution shows better performance in detecting network anomalies. To maintain security in wireless SDN environments, it offers a scalable and flexible solution. The simulation's findings show that while the number of false positives has decreased and detection rates have increased, the system is now suitable for real-time anomaly detection.