For my project, I’m planning to create three machine learning models that perform predictive analytics and act as examples for a series of guidelines I’ll construct for how to create an optimal neural network for predictive analytics. I would really only need the TensorFlow library to create and train the models, Python to use the library, and R to compile and visualize the data used to train the model. Before I start however, I need to have a bevy of background knowledge in various things. First, I would have to have a deep understanding of the structure and common configurations of a neural network. Second, I would have to learn what defines an optimal neural network. Would it just be a tradeoff between accuracy, time, and memory, or is there more to it? Third, I would have to learn the many optimization algorithms that are out there and the math behind them in order to figure out what the best optimization algorithm based on how the data is structured. After I gain the background knowledge, it will be a feasible endeavor. It will just be a matter of collecting the architecture and optimization algorithms from the literature, figuring out the justification of those things, coming up with guidelines, and then testing them out in three scenarios.