On the Level of Precision of a Heterogeneous Transfer Function in a Statistical Neural Network Model
Abstract
A heterogeneous function of the statistical neural network is presented from two transfer functions: symmetric saturated linear and hyperbolic tangent sigmoid. The precision of the derived heterogeneous model over their respective homogeneous forms are established, both at increased sample sizes hidden neurons. Results further show the sensitivity of the heterogeneous model to increase in hidden neurons.
DOI
10.22237/jmasm/1608553560
Included in
Applied Statistics Commons, Social and Behavioral Sciences Commons, Statistical Theory Commons
