Build Regression, Classification, and Clustering Models

In the preceding course, you went through the overall machine learning workflow from start to finish. Now it's time to start digging into the algorithms that make up machine learning. This will help you select the most appropriate algorithm(s) for your own purposes, as well as how best to apply them to solve a problem. A good place to start is with simple linear regression.

The simple model you created earlier works well in many cases, but that doesn't mean it's the optimal approach. Linear regression can be enhanced by the process of regularization, which will often improve the skill of your machine learning model. In addition, an iterative approach to regression can take over where the closed-form solution falls short. In this module, you'll apply both techniques.

Besides linear regression, the other major type of supervised machine learning outcome is classification. To begin with, you'll train some binary classification models using a few different algorithms. Then, you'll train a model to handle cases in which there are multiple ways to classify a data example. Each algorithm may be ideal for solving a certain type of classification problem, so you need to be aware of how they differ.

It's not enough to just train a model you think is best, and then call it a day. Unless you're using a very simple dataset or you get lucky, the default parameters aren't going to give you the best possible model for solving the problem. So, in this module, you'll evaluate your classification models to see how they're performing, then you'll attempt to improve their skill.