Summary

In this chapter, the goal was to get you up and running in the exciting world of neural networks and deep learning. We examined how the methods work, their benefits, and their inherent drawbacks with applications to two different datasets. These techniques work well where complex, nonlinear relationships exist in the data. However, they are highly complex, potentially require a ton of hyper-parameter tuning, are the quintessential black boxes, and are difficult to interpret. We don't know why the self-driving car made a right on red, we just know that it did so properly. I hope you will apply these methods by themselves or supplement other methods in an ensemble modeling fashion. Good luck and good hunting! We will now shift gears to unsupervised learning, starting with clustering.