Difficulty in interpreting recommend() results

User 4540 | 4/13/2016, 2:36:28 AM

Hello, I am trying to interpret the output getting for the below code[data has the userIDs,ItemIDs and Ratings]. Basically I am expecting that it should provide recommendations to all users based on item-ratings. result=graphlab.recommender.itemsimilarityrecommender.create(data,userid='UserId', itemid='ItemId',target='Rating',similarity_type='cosine') reco=result.recommend() print reco

Output: +--------+--------+-------+------+ | UserId | ItemId | score | rank | +--------+--------+-------+------+ +--------+--------+-------+------+ [0 rows x 4 columns]

Can anybody help in explaining the output?


User 19 | 4/13/2016, 3:45:14 AM

Hi schawla,

Is it possible that this user has seen all of the items? By default, the trained model will only recommend items for which no rating exists for the given user.

If that is not the case, can you provide more details about your dataset data?

Thanks, Chris

User 4540 | 4/13/2016, 7:09:25 AM

Yes, I guess that's what happening since I didn't pass any specific user and its related item-ratings, so probably the trained model isn't recommending any item.

And the dataset I have contains list of userId's bunch of items(11) and the Ratings(item ratings rated for each user for each item).