what is neural network sim function?

I set the input layer with 2 neurons and a hidden layer with 3 neurons in Matlab ANN, and trained the model.

Why are these two methods lead to differernt output:

1.

z1 = IW * I' + b1; % Output of the first hidden layer

a1 = logsig(z1); % Apply the sigmoid activation function

z2 = LW1 * a1 + b2; % Final output

output = logsig(z2); % Apply the sigmoid activation function

2.

output = sim(net, I')

I really want to use the first method, but it seems something wrong with it.

Answer 1

[Wang Xiaomeng]

Thank you very much for your help.

Answer 2

[Ayush Aniket]

Hi Wang,

As per my understanding, the neural network analytical equations are not giving you the same result as through “sim” function. The reason behind this is that the "net” object returned by the “feedforwardnet” neural network function in MATLAB uses pre-processing and post-processing functions to process the input and output data.

To achieve the same output through the “sim” function, you can add these lines to your code:

normalized_I' = mapminmax('apply', I', net.inputs{1}.processSettings{1}); %we are applying the same pre-processing as in the feedforwardnet 
output = mapminmax('reverse', output, net.outputs{2}.processSettings{1});%we are applying the same post-processing as in the feedforwardnet 

To read more about the process functions please refer to the following documentation page:

https://www.mathworks.com/help/deeplearning/ug/neural-network-subobject-properties.html#bss4hk6-68

Hope it helps.