Credit analysts generally assess the risk of credit applications based on their previous experience. They frequently employ quantitative methods to this end. Among the methods used...
Abstract—In this work we present a variational formulation for a multilayer perceptron neural network. With this formulation any learning task for the neural network is defined ...
This paper presents a novel architecture for on-chip neural network training using particle swarm optimization (PSO). PSO is an evolutionary optimization algorithm with a growing ...
Amin Farmahini Farahani, Seid Mehdi Fakhraie, Saee...
A flexible neural network (FNN) is a multilayer feedforward neural network with the characteristics of: (1) overlayer connections; (2) variable activation functions for different...
In this paper, a novel and effective criterion based on the estimation of the signal-to-noise-ratio figure (SNRF) is proposed to optimize the number of hidden neurons in neural ne...