Hello.
I have a question regarding the meaning of the variable pol_sit_duration.
Reporting the data dictionary:
Represents how old the current policy characteristics are. […] The pol_sit_duration is a
metric for how up-to-date the characteristics are.
If the number is greater than 1 does it mean that all information about the policyholder are not updated to the current year of that policy?
(eg. pol_sit_duration=3 --> all information in that line are old of three years, so that age of customer are not updated)
Or does it mean “years since the last change in the policy features”?
(eg. pol_sit_duration=3 --> customer has changed something in the policy three years ago, but personal information of him/her are updated, like age)
Your second interpretation is correct. The personal information that can be automatically updated is updated such as age, or the age of the license.
But for example, if a particular policy holder has been with the company for 10 years but updated their vehicle 4 years ago then pol_sit_duration would be 4 while pol_duration would be 10.
Or more commonly, maybe an old policy gains a second driver (say a marriage, or a child learning to drive), then nothing but this second driver feature would updated. That kind of change or demographic shift can be seen from this pol_sit_duration feature.