17.4 Data Cleaning
We have some models now that we can consider decent "starting points." However, we rarely (if ever) deploy the first model we create. That is because there are several types of data cleaning and transformation steps that may improve the model fit even further. You might consider this step as "Round 2" of the data wrangling we performed earlier. In particular, we will try the following:
Math transformations
Changing data type
Managing outliers