作者: Yan Zhao , Qingshan Liu
DOI: 10.1016/J.MATCOM.2012.03.013
关键词:
摘要: In this paper, a generalized recurrent neural network is proposed for solving @e-insensitive support vector regression (@e-ISVR). The @e-ISVR first formulated as convex non-smooth programming problem, and then generalize with lower model complexity designed training the machine. Furthermore, simulation results are given to demonstrate effectiveness performance of network.