neuron for a back prop neural net (BPnnet) More...
#include <BPnnet/BPneuron.H>
Public Member Functions | |
BPneuron () | |
Constructor. | |
~BPneuron () | |
Destructor. | |
void | assignInput (double input) |
Assign given weighted input sum [S_j] and activation level. | |
double | calcOutputDelta (const double target) |
Calculate and set delta for output neurons. | |
void | calcHiddenDelta (double weightedDeltaSum) |
Calculate and set delta for hidden neurons. | |
double | getDelta () |
Return delta (error level). | |
double | getActivationLevel () |
Return activation level. |
neuron for a back prop neural net (BPnnet)
Definition at line 42 of file BPneuron.H.
BPneuron::BPneuron | ( | void | ) |
Constructor.
Definition at line 42 of file BPneuron.C.
BPneuron::~BPneuron | ( | ) |
Destructor.
Definition at line 50 of file BPneuron.C.
void BPneuron::assignInput | ( | double | input | ) |
Assign given weighted input sum [S_j] and activation level.
Activation Level a_j = f(S_j)
Definition at line 54 of file BPneuron.C.
void BPneuron::calcHiddenDelta | ( | double | weightedDeltaSum | ) |
Calculate and set delta for hidden neurons.
For hidden layer neurons; compute hidden delta from weighted sum of output deltas [delta_j = sum_k(delta_k * w_kj)*f'(S_j)] returns delta_j
Definition at line 70 of file BPneuron.C.
double BPneuron::calcOutputDelta | ( | const double | target | ) |
Calculate and set delta for output neurons.
For output layer neurons; finds difference between output and target and multiplies by derivative of sigmoid function [delta_j = (t_j - a_j)*f'(S_j)] returns the difference [t_j - a_j] for RMS error calculation
Definition at line 61 of file BPneuron.C.
double BPneuron::getActivationLevel | ( | void | ) |
Return activation level.
Definition at line 83 of file BPneuron.C.
double BPneuron::getDelta | ( | void | ) |
Return delta (error level).
Definition at line 77 of file BPneuron.C.