Summary

In this chapter, we briefly discussed the learning process for machines, which, to some extent, mimics that of human beings. We described how a model, which is a simplified representation of the problem that we want to solve, can be used to apply machine learning to find a solution.

Using a linear regression model, we built a simple supervised predictive model and explained how to use it. We then discussed the difference between regression and classification, and showed the properties of the input variables and features.

Underfitting and overfitting are two of the main concerns when training a machine learning model. We explained what they are and suggested methods to avoid them.

Finally, different types of target variables require different algorithms and evaluation methods to test the quality of the model – we discussed this in detail in the final sections.

In the next chapter, we are going to solve some real-life problems using machine learning and explore how some supervised and unsupervised models are built.