On the approximation by weighted ridge functions

2007/08/24 by Ismailov, Vugar
#41A30 #41A50 #41A63 #Classical Analysis and ODEs (math.CA) #FOS: Mathematics

paper · doi:10.48550/arxiv.0708.3356

Abstract

We characterize the best L₂ approximation to a multivariate function by linear combinations of ridge functions multiplied by some fixed weight functions. In the special case when the weight functions are constants, we propose explicit formulas for both the best approximation and approximation error.

Related

• On some aspects of approximation of ridge functions
• Approximation by Combinations of ReLU and Squared ReLU Ridge Functions\n with \ℓ¹ and \ℓ⁰ Controls
• Notes on ridge functions and neural networks
• A note on the equioscillation theorem for best ridge function approximation
• On best approximation by multivariate ridge functions with applications to generalized translation networks