recommendation

Type Members

1. class GlintFMPair extends Estimator[GlintFMPairModel] with GlintFMPairParams with DefaultParamsWritable

   

   Distributed pairwise factorization machine / LightFM.

   Distributed pairwise factorization machine / LightFM.

   Pairwise factorization machines are trained on implicit-feedback training instances to rank all items with observed user-item training instances above all other items for the user, using bayesian personalized ranking.

   This is an implementation using Glint parameter servers with custom methods for network-efficient training. A Spark application with the parameter servers has to be started beforehand and the host of the parameter server master given as parameter to this implementation.

2. class GlintFMPairModel extends Model[GlintFMPairModel] with GlintFMPairParams with MLWritable

   

   Model fitted by GlintFMPair.

   Model fitted by GlintFMPair.

   For simplicity, this implementation uses the parameter servers for recommendation. Real use cases will require a different implementation which exports the linear weights and the latent factors, uses approaches like locality-sensitive hashing for recommendation in sublinear time and does not use parameter servers at recommendation time.

3. class KNN extends KNNRecommender

   

   Collaborative filtering through user-based k-nearest neighbours.

4. class PopRank extends Estimator[PopRankModel] with PopRankParams with DefaultParamsWritable

   

   Popularity ranking recommender.

5. class PopRankModel extends Model[PopRankModel] with PopRankParams with MLWritable

   

   Model fitted by PopRank.

6. class ResizeVector extends UDF2[SparseVector, SparseVector, SparseVector]

   

   UDF to resize a sparse vector to the same size as another sparse vector

7. class SAGH extends Estimator[SAGHModel] with SAGHParams with DefaultParamsWritable

   

   Same-Artist Greatest Hits recommender.

8. class SAGHModel extends Model[SAGHModel] with SAGHParams with MLWritable

   

   Model fitted by SAGH.

9. class TfIdfKNN extends Estimator[TfIdfKNNModel] with _TfIdfKNN with DefaultParamsWritable

   

   Collaborative filtering through user-based k-nearest neighbours, using TF-IDF weighting for user similarity.

10. class TfIdfKNNModel extends Model[TfIdfKNNModel] with _TfIdfKNN with KNNRecommender with MLWritable

    

    Model fitted by TfIdfKNN.

11. class WeighVector extends UDF2[SparseVector, Double, SparseVector]

    

    UDF to multiply a sparse vector by a weight and divide it by the number of values