Orthogonal kernel matching pursuit (OKMP) for constructing sparse kernel models has been recently introduced

in which a greedy scheme is utilized to select a single element per iteration.To improve the efficiency and performance of the greedy-scheme-based OKMP

a greedier algorithm is considered.The main contribution is the development of a new selection strategy that effectively selects several elements in each iteration.The efficiency is achieved by reducing the regressor steps

thus the computation time of the orthogonalization that each newly selected regressors to all the selected terms before is saved.A pruned algorithm is proposed based on the similarity of the atoms to improve the accuracy of the approximation.Numerical results and computational complexity analysis show that this new scheme is capable of producing a much sparser regression model with better generalization than the conventional approaches.