I have few questions concerning Neural networks especially the Neural Network Toolbox and i'd really appreciate it if you could give me some answers. My questions are as follows:
1/ How can we get the weight matrices used in the training, testing and Validation sequences?
As far as i know, the final weights of the training sequence are used to test the model in the testing sequence using the same learning algorithm and a non used database and to validate the model, we use the final testing weights with the same learning algorithm.
2/ how do exactly the testing and validation sequences work in Matlab?
3/ Are also the training biases used in the testing and validation sequences?
4/ Why can't we have more than one hidden layer in the multilayer perceptrons problems? ( for example fitting problems)
Thank you in advance
Best Answer