MATLAB: Too many input arguments in class method for neural network.

arrayinputMATLABneural network

My function takes in a single input and provides a single output, but of the array type. Class properties are defined properly earlier and this as been tested to work (e.g. net.Num_Layers) Sizes is an array denoting number of neurons in each layer (e.g. net.Sizes = [3 4 2]) When trying to test individual method, a Network object is created in command line.
 >>net = Network(sizes)
   net = 
        Network with properties:
         Sizes: [3 4 2]
    Num_Layers: 3
       Weights: [4×3×4 double]
        Biases: [4×3 double]
>> a = [1 2 3]';
>> net.feedforward(a)
Error using Network/feedforward
Too many input arguments.
Here a snippet of my feed-forward method for the Network class:
       function [a] = feedforward(a)
            %return output of network if column vector "a" is input.
            for i=2:net.Num_Layers
                w = squeeze(net.Weights(1:net.Sizes(i),i,1:net.Sizes(i-1)));
                b =net.Biases(1:net.Sizes(i),i);
                a = w*a-b;
            end
        end
And here is my constructor method for reference if its relevant:
          function net = Network(sizes)
%           The list ``sizes`` contains the number of neurons in the
%         respective layers of the network.  For example, if the list was
%         [2, 3, 1] then it would be a three-layer network, with the first
%         layer containing 2 neurons, the second layer 3 neurons, and the
%         third layer 1 neuron.  The biases and weights for the network are
%         initialized randomly, using a Gaussian distribution with mean 0,
%         and variance 1.  Note that the first layer is assumed to be an
%         input layer, and by convention we won't set any biases for those
%         neurons, since biases are only ever used in computing the outputs
%         from later layers.
            if nargin == 0
                net.Sizes = 0;
            end
            net.Num_Layers = length(sizes);
            net.Sizes = sizes;
            %Intialize and pad with zeros
            net.Biases=zeros(max(sizes),net.Num_Layers);
            net.Weights=zeros(max(sizes),net.Num_Layers,max(sizes));
            for i=2:net.Num_Layers
                net.Biases(1:sizes(i),i) = randn(sizes(i),1);
            end
            %layers are stored along columns i
            %connecting from previous layer along rows j
            %connecting to current layer along pages k
            for i=2:(net.Num_Layers)
                net.Weights(1:sizes(i),i,1:sizes(i-1)) = ... 
                    randn(sizes(i),1,sizes(i-1));
            end    
        end

Best Answer

A class method must have its class object explicitly in its argument list. So the following will work
function [a] = feedforward(obj, a)
 %return output of network if column vector "a" is input.
 for i=2:net.Num_Layers
   w = squeeze(net.Weights(1:net.Sizes(i),i,1:net.Sizes(i-1)));
   b =net.Biases(1:net.Sizes(i),i);
   a = w*a-b;
 end
end
Related Question