Sparse Approximation Through Boosting for Learning Large Scale Kernel Machines | IEEE Journals & Magazine | IEEE Xplore