I have the below information for around 1000 stores. enter image description here

The columns competition distance, competitionOpenSince , Promo2Since columns have NA's.

1.Promo2SinceWeek/Year is missing only for stores that did not run the promotion(Promo2=0).

2.Competition related columns are missing in random.

All these columns are specific to a store. Different stores will have different values. replacing a missing value by taking the mean/median of all other stores does not make sense to me.

I want to fit a linear regression/ regression trees model to predict sales. I really need some expert suggestions/ideas to handle these NA's with the appropriate method.

If required, I am ready to add the dataset here


Essentially you are looking for ways to impute the missing values.

Please take a look at the following article which gives you a few different strategies for missing values. MICE works well in many cases, but it depends on your particular use case.

| cite | improve this answer | |

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.