Artificial Intelligence for Business, 2nd Edition

13

Artificial Neural Networks in Action

In this chapter:

• Leveraging the power of data

• Putting hidden layers to work

• Going behind the scenes with activation functions

• Tweaking inputs with relative weights

• Tweaking a neuron’s output with bias

You’ve seen that machine learning (ML) is an extremely robust form of pattern matching. You can use ML algorithms to find complex relationships or even classify data based on patterns that defy human perception. Neural networks take this to the next level. Here you can use thousands or even millions of artificial neurons to analyze data and identify subtle patterns.

Let’s look at a fairly common ML classification ...