Neural Network Training Based on Kolmogorov's Theorem

This neural network architecture implements Kolmogorov's theorem with 14 input nodes, 29 hidden layer nodes using tansig transfer functions, and 8 output nodes using logsig transfer functions, trained with the traingdx algorithm.