I have a dataset of customers:

age  height  weight eye_colour
30    174    74       Nan
20    191    71       Nan
28    165    56       Brown

I would like to identify similarities between customers where the eye_colour is Nan in order to help identify the cause of this missing value.

Is there any method to help identify similarities?


In general identifying similarities is done with clustering, but in this case what you're looking for is a potential pattern in the data leading to a missing value in a specific column, right? So I would try to train a decision tree using all the columns except the target one as features, and using a binary class indicating whether the target column has a missing value or not. Visualizing the decision tree obtained after training should show the pattern if there is one.

  • $\begingroup$ Thank you Erwan, I will try your method and inform you :) $\endgroup$ Jun 17 '19 at 8:23

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.