In the company I work for there are 2 different evaluation metrics for a song:
- Yes / No (Equivalent to like/dislike)
- 1-5 Scale
Customers can use both to rank songs they like. I would like to create a model that predicts the next possible songs you would like. Currently, I'm ignoring the Binary data. I wonder if there's a good way of utilizing the Binary data as tagged data [And not as a feature].
I've thought about two possible solutions:
- Calculate the 1-5 Scale rating of each user and then take the (mean + std) as 'yes' and (mean - std) as 'no'.
- Calculate the percentile of 1-5 Scale rating of each user and then take p75 as 'yes' and p25 as 'no'.
In case user doesn't have 1-5 scales I simply ignore him.
I guess that are better ways? (Maybe more empirically correct ways?)