Explaining Null Values vs Missing Values to a Predictive model

r
missing_values

#1

while predicting employee churn I want to include Termination Age. But for employees who are still active the termination age will not be applicable. So this is a case of null value and not missing value. How do we treat these values in my model? I definitely know that we cant impute, at the same time this seems to be an important predictor.


#2

Question: Is termination age the metric of an employee’s age at which you should fire an employee?


#3

Termination age is not a status or classification. So you can still calculate their current staying age. For active employee you call it stay age and for churned employees you can call this termination age.

It is simply the time at the company. It will be a continuous variable and it all comes down to interpretation.


#4

Along with the solutions provided by @daveeed @vivekps , I would like to add that you can give a separate value for showing the model that it is null. For example, giving a value such as “-1” would explain the model that it is different for usual values