1
$\begingroup$

I have been working on a supervised ML use case where dataset has Numerical (Price), Categorical(Category) and Textual data(Description) as features. Description feature has about 30% missing values. I don’t want to drop them as data set is small and it would cause information loss. Looking for any suggestion on how to handle missing value ?

$\endgroup$

1 Answer 1

0
$\begingroup$

A "missing text" can be considered simply an empty text, and an empty text is still a valid text: as long as the design of the code doesn't assume non-empty text, it should not cause any issue. In particular it can be consistently encoded with bag of words representations.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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