Model Building - II in Linear Regression

Let's now look at the next steps in model building.

Upon the introduction of more variables, the model starts becoming complex. Some of the variables, which were significant earlier, now become insignificant. Variables that have better alternatives are replaced.

You brought back variables that were previously removed to test whether they fit the business requirement better. For example, you brought back Visitors as a predictor against Views to the platform because Visitors could be impacted through marketing actions.

In the latest version of the model (Model 6), Ujjyaini spotted two major problems:

1. The variable ‘Visitors’ has become insignificant with a large p-value, and
2. The variable ‘character A’ has changed its sign.