Neural network

Write a python program setting up the neural network for this case study, the network should have the structure described in the book, including initial layer sizes, activation functions, and weight matrix initializations.
Extend the dataset as described, by duplicating the healthy records until the number of healthy records equals the unhealthy records.
All of this should be in a loop to do a new train test split with each run but the network structure does not need to change with each run.
Compare your results to that described in the book.
You are to turn in an .ipynb file with your code.For this case study, document the changes that you are making and the effect they have on the network (in terms of the accuracy and loss). Tell me what the final (best) structure and initializations you would recommend to someone trying to duplicate this study.