Public Member Functions

Weights_Initializer Class Reference

Implements the following algorithm: "Improving the Learning Speed of 2-Layer Neural Networks by Choosing Initial Values of the Adaptive Weights" by Derrick Nguyen and Bernard Widrow. More...

#include <Weights_Initializer.h>

List of all members.

Public Member Functions

 Weights_Initializer (int rnd_seed=-1)
 Constructor, takes internal random numbers generator seed value as a parameter. If no seed is specified _time32(NULL) function result is taken as a seed.
void Set_Weights_Count (int count)
 Sets the number of weights to calculate.
void Set_Hidden_Neurons_Count (int count)
 Sets the number of neurons in hidden layer.
void Get_Weights (std::vector< double > &output, int weights_count=0, int hidden_neurons_count=0)
 Calculates initial weights based on weights_count and hidden_neurons_count.
double Get_Bias_Weight ()
 Calculates random bias weight based on weights_count and hidden_neurons_count.
bool Validate (const std::vector< double > &input, double precision=0.0001)
 Validates Get_Weights() result with the specified precision.

Detailed Description

Implements the following algorithm: "Improving the Learning Speed of 2-Layer Neural Networks by Choosing Initial Values of the Adaptive Weights" by Derrick Nguyen and Bernard Widrow.

MCPN Logo
kr0st © 2010