"lol"
Not easy maths....
.
delta weight adjustment = learning rate * Original activation * difference between expected and actual output
@korrelan thanks very much.
i think that this one iused for the forward propergation rules (only on the hidden layer?) i used it on the input layer (silly me) im wondering if it makes a difference...
i will experiment...
Ive done the forward Ann the same as that formula.. and the backward as the derivatives... I only have derivative formulaes for the sigmoid and the Hyperbolic tangent, when drawn they look really simular in shape. so even these activation functions im sure are the same...
im going to try figure out the gausian derivative... among the others.. although some do not have derivatives.
my simple tests are just the times table... inputs 1,2,3,4,5,6,7,8,9,10 outputs 2,4,6,8,10,12,14,16,18,20
then see if it can do the three times table...
I cracked it today (calculating the deltas) for each layer then calculating the errors for each layer and adjusting the weights relative to thier respctive errors for each node in each layer. (real Tricky)
Private Function ForwardLearning(ByRef NN As NeuralNetwork) As NeuralNetwork
'ApplyLearning rule
'inputLayer?
'Weight = Weight LocalError * LearningRate * input
NN.Params.BiasedTerm = NN.Params.BiasedTerm + NN.Params.LearningRate * localError
For Each Node As Neuron In NN.InputLayer.Nodes
Node.weight = Node.weight + localError * NN.Params.LearningRate * Node.input
Next
For Each Node As Neuron In NN.InputLayer.Nodes
Node.weight = Node.weight + localError * NN.Params.LearningRate * Node.input
Next
For Each Node As Neuron In NN.HiddenLayer.Nodes
Node.weight = Node.weight + localError * NN.Params.LearningRate * Node.input
Next
Return NN
End Function
Private Function BackwardLearning(ByRef NN As NeuralNetwork) As NeuralNetwork
Dim deltaOutput As Double = 0.0
Dim DeltaInput As Double = 0.0
Dim deltaHidden As Double = 0.0
Dim deltaOutputError As Double = 0.0
Dim deltaHiddenError As Double = 0.0
For Each Node As Neuron In NN.OutputLayer.Nodes
'LearningRule
'Di = LocalError * SigmoidDerivitive(Node.input)
deltaOutput = Node.NeuronError * TransferFunctions.EvaluateTransferFunctionDerivative(Node.input, NN.Params.FunctionType)
'nw=w+lr*node.output*Di
Node.weight = Node.weight + NN.Params.LearningRate * Node.output * deltaOutput
deltaOutputError = deltaOutputError + deltaOutput
Next
For Each Node As Neuron In NN.HiddenLayer.Nodes
'dj = SigmoidDerivitive(Node.input) * SumOfDeltaOutput*weights
deltaHidden = TransferFunctions.EvaluateTransferFunctionDerivative(Node.input, NN.Params.FunctionType) * SumWeights(NN.OutputLayer) * deltaOutputError
Node.weight = Node.weight + NN.Params.LearningRate * Node.output * deltaHidden
deltaHiddenError = deltaHiddenError + deltaHidden
Next
For Each node As Neuron In NN.InputLayer.Nodes
DeltaInput = TransferFunctions.EvaluateTransferFunctionDerivative(node.input, NN.Params.FunctionType) * SumWeights(NN.HiddenLayer) * deltaHiddenError
node.weight = node.weight + NN.Params.LearningRate * node.output * DeltaInput
Next
Return NN
thanks guys ..... im so close its biting me in the "Ann"
