2
$\begingroup$

I have data similar to what you see in the picture. I want to use a RandomForest Regression model where I can use fields (excluding MONTH_END_DT and LOCATION_ID) to predict REVENUE_PER_UNIT. The idea/goal being that, if I know how many SERVICEABLE_UNIT_CNT I have, and the number of HAPPY, NEUTRAL, or ANGRY customers I have, I can predict the REVENUE_PER_UNIT.

My question is: considering I don't want to do a time-series analysis, would I drop the duplicates in this dataset (notice the highlighted fields are duplicates if you remove location and month end)? example data

$\endgroup$
1
  • $\begingroup$ Hi Larry, Welcome to the community. Please consider upvoting/marking the answer as correct if you find anything useful on your post. $\endgroup$
    – Kriti
    Jan 4 at 19:32

2 Answers 2

1
$\begingroup$

Yes, you should be removing the duplicates considering you would be dropping the month_end_dt and location_id column as well before feeding the data to the model.

$\endgroup$
1
  • $\begingroup$ Thank's @Kriti, my points are too low to upvote your answer, but I accepted it. $\endgroup$ Jan 4 at 20:10
1
$\begingroup$

I would consider not removing the duplicates, instead some light feature engineering by extracting the month of the year as a qualitative feature, as it's possible customer behaviour and sentiment is influence by given months.

$\endgroup$

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.