0
$\begingroup$

Given a house sales dataset with number of houses sold each day and their attributes (i.e., price, number of rooms, size, etc.) - if on a specific day there's a spike/drop in sales, what are some techniques to automatically identify in which segments/combination of sub-segments were these drops? for example: the drop was because there were a lot less top priced, large houses sold.

Also, is there a technical name for this problem? I'd like to search and read more about it

$\endgroup$
0
$\begingroup$

Hello you may start with reading Contextual Outlier Interpretation. You may also adapt LIME for the scenario of outlier detection. The key idea is to assume that you can model the local decision boundary by a linear model. Then you are able to interpret influence of variables on the outlier score.

$\endgroup$

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.