Summary

In this chapter, we started out with laying the process for building a machine learning workflow, starting from designing the problem and moving to deploying the model. We briefly discussed simple and multiple and logistic regressions along with all the evaluation metrics needed to interpret and judge the performance of the model. These two algorithms demonstrate the supervised learning for regression and classification problems, respectively.

Throughout the chapter, we used the Beijing PM2.5 dataset to build the models. In the process, we also converted a regression problem to a classification problem by simply re-engineering the dependent variable. Such re-engineering is often taken up on real-world problems to suit a particular use case.

In the next chapter, we will delve into the details of regression algorithms and will elaborate the various types of regression algorithms beyond linear regression and discuss when to use which one.