Identification Of Malicious User To Combat Ssdf Using Trust Based Machine Learning Techniques In Cognitive Radio Networks

Abstract

Spectrum Sensing data falsification (SSDF) attack is one of the major threats in cognitive radio networks. In this paper a neural network (NN) machine learning (ML) is trained with multifactor trust based computed from the spectrum sensing output of the SU’s to identifythe Malicious Users (MU’s) in the SSDF attack.The performance of NN, Support Vector Machine (SVM), Naïve Bayes (NB) and Logistic Regression (LR) are compared. Also the performance of all the mentioned classifiers are validated with ‘k’- fold cross validation method. Simulations have proved that the logistic regression ML is the best model to evaluate the trust based data set with 100% accuracy.