A Supervised Learning Approach for Differential Entropy Feature-based Spectrum Sensing

In this work, we consider a supervised machine learning-based approach for spectrum sensing in cognitive radios. The noise process is assumed to follow a generalized Gaussian distribution, which is of practical relevance. For classification, we consider the differential entropy estimate in the received observations as a feature vector. For our comparative study, we consider the support vector machine, K-nearest neighbor, random forest and logistic regression techniques. Through experimental results based on real-world captured datasets, we show that the proposed differential entropy feature-based technique outperforms the energy-based approach in terms of probability of detection. The proposed technique is particularly useful under low signal-to-noise ratio conditions, and when the noise distribution has heavier tails.