Outputs of the second layer in PNN neural network

How to get the output response of the second layer in PNN network when use the command : newpnn in matlab.

Answer 1

[Snehal]

Hello,

In a Probabilistic Neural Network (PNN) created using MATLAB's ‘newpnn’ function, the layers are structured as follows:

1. Input Layer: Passes the input features directly to the pattern layer without any computation
2. Pattern Layer (First Computational Layer): Computes similarity scores between the input and each training sample using a radial basis function
3. Summation Layer (Second Computational Layer): Aggregates the outputs from the pattern layer for each class
4. Decision Layer (Output Layer): Determines the class with the highest aggregated score

The Summation Layer is considered the second computational layer in a PNN, and my understanding is that you want to extract the output responses for this layer. You can use the following code snippet for implementing this:

% Let’s assume that you have already created a PNN using the following command:
% net = newpnn(P,T,spread);
% where P: matrix of input vectors
% T : matrix of target class vectors
% spread: Spread of radial basis functions 

% the syntax that you can use to get the output responses is:
patternLayerOutput = radbas(dist(net.IW{1,1}, inputSample) * spread);
summationLayerOutput = net.LW{2,1} * patternLayerOutput;

Additionally, you may refer to the following documentation links for more details:

• https://www.mathworks.com/help/deeplearning/ug/probabilistic-neural-networks.html
• https://www.mathworks.com/help/deeplearning/ref/newpnn.html

Hope this helps!