Data Science Programming All-in-One For Dummies

Chapter 2

Building Deep Learning Models

IN THIS CHAPTER

Understanding neural network basics

Getting deeper into neural networks

Applying what you know to deep learning

The idea of ensembles of learners appears in the previous chapter of this minibook. To make the computer better able to model complex real-world problems, you combine algorithms in different ways. Each algorithm adds to the whole. The computer doesn’t actually understand anything. You rely on math to create a model that approximates learning. Creating models that approximate learning is what this chapter is about, too, but now you move to another level of learning called deep learning. In deep learning, a computer builds a complex structure called a neural network that is able to delve into datasets at an incredibly low level and model the data more precisely than any ensemble. The whole principle relies on mimicking the human brain using a mathematical neuron.

The first part of the chapter discusses the nature of a computer neuron and tells why it’s important. However, it starts with an historical view of the impact on data science by the perceptron, a device that was amazing in its time, but also oversold.

After ...