Matchbox Recommender Azure ML Studio predicting same items RRS feed

  • Question

  • Hi!

    I'm trying to deploy a recommendation sys for articles.

    My training data to the recommender includes:

    rating triplets: userId, articleId, Ratings

    User features: userid, dietaryType pref

    Article features(item):ArticleId, category

    Score Matchbox Recommender Kind: Item recommendations

    Recommended item selection: from all users

    During prediction time the models predicts few specific items for all users, even though these articles are not even rated high.

    Any suggestion/explanation will be appreciated.



    Thursday, August 22, 2019 9:00 AM

All replies