作者: Illya Kokshenev , Antonio Padua Braga
DOI: 10.1016/J.NEUCOM.2007.11.021
关键词:
摘要: The problem of inductive supervised learning is discussed in this paper within the context multi-objective (MOBJ) optimization. smoothness-based apparent (effective) complexity measure for RBF networks considered. For specific case network, bounds on are formally described. As synthetic and real-world data experiments show, proposed MOBJ method capable efficient generalization control along with network size reduction.