Context aware recommender system

#1

Can i use fastai for context aware recommender system problem?

I have this dataset https://github.com/irecsys/CARSKit/blob/master/context-aware_data_sets/Music_InCarMusic.zip

So like in the lesson 4 we have an example of collaborative filtering with userid and movieid and rating for each combination.
However in the ‘context-aware’ dataset above. Which is for music recommendation depending on context i have more ‘features’ like: userid, itemid, rating, drivingstyle, landscape, mood, naturalphenomena, roadtype, sleepiness, trafficcondition and weather.

For example user A may give score 4 to song X in the morning but socre 2 for the same song in the evening, etc.

Is fasti supporting such a problem?

Thanks!

0 Likes

(Jeremy Easterbrook) #2

I have been curious of passing categorical variables as well to some collaborative filtering.

0 Likes