层次径向基神经网络的全局逼近理论

UNIVERSAL APPROXIMATION THEORY OF HIERARCHICAL RADIAL BASIS FUNCTION NETWORK

摘要: 文中提出了一种层次径向基神经网络ＨＲＢＦＮ（ｈｉｅｒａｒｃｈｉｃａｌｒａｄｉａｌｂａｓｉｓｆｕｎｃｔｉｏｎｎｅｔｗｏｒｋ），并且证明了ＨＲＢＦＮ是一个全局逼近器．ＨＲＢＦＮ更适合于具有层次结构的应用领域，因而具有极为重要的理论与应用意义

Abstract: In this paper, a new hierarchical radial basis function network HRBFN is presented, and its universal approximation property is proved. HRBFN is very suitable for application domains with hierarchical structures, and therefores has very important theoretical and practical significance.