Practical Predictive Analytics

Transformed variables

Transformed variables are variables that you create which did not exist in the raw data. Here are some examples of how you can create transformed variables:

• You can bin a numeric variable into several distinct categories such as High (all numeric values greater than 10) or low or any value less than 10. Notice that binning results in losing information, but allows you some flexibility by being able to name things and promote understandability
• Transforming count data into percentages by dividing a cell count by the total population.
• Standardizing the data in the modeling phase; it is often useful to work with standardized variables, as opposed to using the raw data itself. A standardized variable is a transformation ...