Weighting informativeness of bag-of-visual-words by Kernel optimization for video concept detection