Summary

To produce a predictive model with linear regression, we have to be sure that the variables involved have a strong relationship.

Linear regression is a supervised machine learning algorithm because it needs known data for training and testing. Even after the model is in production, we need to keep checking whether the model needs an update with new data.

Before we design a model, we have to use statistical methods to probe the relationship between the variables. These methods include the coefficients of correlation and regression, t-statistics, and p-values. The machine learning models are defined and trained on a portion of the data. Then, we test the model with the remaining data, and finally, use the model to make predictions. We have to use our judgment and experience to decide whether the model is accurate or not. Build a chart with predictive values from the model to see whether the values make sense based on your experience.

In the next chapter, we will be...