一种新的RBF网络两级学习设计方法

为了简化径向基网络结构,构造出良好泛化性能力的网络,提出了一种径向基(RBF)网络的两级学习新设计方法.在下级将正交最小二乘法(OLS)与A-最优设计方法(A-opt)相结合(OLS+A-opt),引入一种基于A-最优设计准则的混合代价函数,同时优化网络模型的逼近性能及模型的充分性,自动构建结构节俭的RBF网络模型;而方法中的关键学习参数A-最优代价系数通过上级粒子群优化方法(PSO)优化获取最佳值.仿真结果表明该方法所设计的RBF网络不仅具有较好的泛化性能,而且也具有良好的模型鲁棒性及充分性,是一种有效的RBF网络设计方法.

A New Two-level Learning Design Approach for Radial Basis Function Neural Network

A new two-level learning algorithm for designing radial basis function(RBF) networks is proposed in this paper in order to simplify the structure of RBF networks and obtain better generalization ability.This new learning algorithm combined orthogonal least squares(OLS) with A-optimality design method(A-opt),and introduced a composite cost function based on A-optimality design criterion,which simultaneously optimized the model approximation ability and model adequacy,and constructed a parsimonious model size...