Boundedness and Convergence of Online Gradient Method With Penalty for Feedforward Neural Networks | IEEE Journals & Magazine | IEEE Xplore