RVM classification of hyperspectral image based on wavelet Kernel function

Relevance Vector Machine (RVM) technique as a new machine learning method is illustrated in details. It is a novel kind of learning method which is based on Bayesian learning theory. RVM presents the good generalization performance, and its predictions are probabilistic. Relevance vector machine mathematics model doesn´t have regularization coefficient and its kernel functions do not need to satisfy Mercer´s condition. The wavelet kernel is a kind of multidimensional wavelet function that can approximate arbitrary nonlinear functions. The existence of wavelet kernels is proven by results of theoretic analysis. Computer simulations show the feasibility and validity of wavelet relevance vector machines in hyperspectral image classification.