I am currently doing a linear regression, where i try to predict the housing prices based on different variables that describe the house's spatial features (such as the distance to the closest city, the closest road etc.).

My problem is, one of the original datasets have only calculated the distance to the closest road if the road is within a 2 km radius of the house. so any house which do not have a road closer than 2 km to it have gotten a NULL value instead of the distance.

I was therefore wondering, is it possible to replace these null values? for example with some value above 2 km?

  • $\begingroup$ This will be tricky... First I need you to explain why your profile picture is Thyra Frank $\endgroup$ – Repmat Dec 5 '17 at 14:19
  • 2
    $\begingroup$ There is a tag on this site interval-censoring which contains some threads which may be of interest to you. $\endgroup$ – mdewey Dec 5 '17 at 16:45

My suggestion would be to include a dummy if the values are missing. If it is significant you can conclude that living more than 2km from a road (decreases, probably) the value of the house.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.