how to calculate the output of neural network manually using input data and weights.

i am having ann program with 3 inputs and one output. i am using back propagation and feed forward network. the activation functions are tansig and purelin. no of layer is 2 and no of neuron in hidden layer is 20. i want to calculate the output of network manually using the input and weights(iw,lw,b) i need an equation to find the output. can you help me?

[prabakaran jayaraman]

thanks greg

Answer 1

[Greg Heath]

When I-dimensional "I"nput x and O-dimensional "O"utput target t are normalized via the default mapminmax (or mapstd),the relationship between the normalized input and output is

yn = repmat( b2, O, N ) + LW * tanh( repmat( b1 , I, N ) + IW * xn);

Thank you for formally accepting my answer

Greg

[prabakaran jayaraman]

thanks greg. how to get ∑ (Xi * IW). if i am having 3 inputs and iw as 20 X 3 matrix

[Greg Heath]

IW does not act on the original weights. It acts on the normalized weights. The default normalization documentation is

 help mapminmax
 doc mapminmax.

Search for examples using a subset of

 greg xsettings tsettings

Greg

Answer 2

[Amir Qolami]

This works for any number of hidden layers and neurons;

function output = NET(net,inputs)

    w = cellfun(@transpose,[net.IW{1},net.LW(2:size(net.LW,1)+1:end)],'UniformOutput',false);
    b = cellfun(@transpose,net.b','UniformOutput',false);
    tf = cellfun(@(x)x.transferFcn,net.layers','UniformOutput',false);

    %%mapminmax on inputs
    if strcmp(net.Inputs{1}.processFcns{:},'mapminmax')
        xoffset = net.Inputs{1}.processSettings{1}.xoffset;
        gain = net.Inputs{1}.processSettings{1}.gain;
        ymin = net.Inputs{1}.processSettings{1}.ymin;
        In0 = bsxfun(@plus,bsxfun(@times,bsxfun(@minus,inputs,xoffset),gain),ymin);
    else
        In0 = inputs;
    end
    In = cell(1,length(w));     Out = In;
    In{1} = In0'*w{1}+b{1};
    Out{1} = eval([tf{1},'(In{1})']);
    for i=2:length(w)
        In{i} = Out{i-1}*w{i}+b{i};
        Out{i} = eval([tf{i},'(In{',num2str(i),'})']);
    end

    %%reverse mapminmax on outputs
    if strcmp(net.Outputs{end}.processFcns{:},'mapminmax')
        gain = net.outputs{end}.processSettings{:}.gain;
        ymin = net.outputs{end}.processSettings{:}.ymin;
        xoffset = net.outputs{end}.processSettings{:}.xoffset;
        output = bsxfun(@plus,bsxfun(@rdivide,bsxfun(@minus,Out{end},ymin),gain),xoffset);
    else
        output = Out{end};
    end

end