Machine Learning in Cognitive Radios
Sudharman K. Jayaweera · 2014
This chapter focuses on identification of value of learning ability to a cognitive radio and investigation of various machine learning algorithms and their uses in achieving the objectives of a cognitive radio. One of the commonly used learning frameworks in various application contexts is the artificial neural networks (ANNs). In developing ANNs, one observes that biological neurons are best modeled as being nonlinear decision devices. Support vector machines (SVM) addresses the representation issue by first non-linearly transforming the input data into a higher-dimensional space and then designing linear decision boundaries to classify this transformed data. Reinforcement learning is essentially learning by trial and error. The most widely investigated multiagent learning paradigm is based on (noncooperative) game theory. Game theory provides a systematic way to model interactions among rational agents.