Machine Learning with Swift

Where to use GD and normal equation

If the goal is just to add linear regression to your app, at this point you are done; however, there is another more interesting way to obtain the same coefficients—an optimization technique known as gradient descent. The gradient descent algorithm and its multiple descendants are used to find loss function minimums in many machine learning algorithms, including deep neural networks, where a direct analytical solution, like with linear regression is not possible. So, we'd better try it on a simple example like linear regression, so we're already familiar with it when we discuss more complex algorithms.