For deep learning you may need to follow: https://machinelearningmastery.com/how-to-develop-convolutional-neural-network-models-for-time-series-forecasting/

Via Matlab:

https://www.mathworks.com/products/deep-learning.html

Muhammad Ali

Dear Brother Ali

Thanks for your replies but I am working on shallow network and I want a network to be designed on the following matrix and elements of the matrix will act as a weight between two layers like 1st row will become weight between the input layer and 1st Hidden layer, 2nd row will act as weights between 1st hidden layer and 2nd hidden layer, and so on but the last hidden layer will be fully connected to output layer.

I checked the links you sent but it didn't solved my purpose.

Of course it's possible-just specify the elements of the matrices mentioned, so they couple the nodes in the way desired.