Implicit factorization models¶
Factorization models for implicit feedback problems.

class
spotlight.factorization.implicit.
ImplicitFactorizationModel
(loss='pointwise', embedding_dim=32, n_iter=10, batch_size=256, l2=0.0, learning_rate=0.01, optimizer_func=None, use_cuda=False, representation=None, sparse=False, random_state=None, num_negative_samples=5)[source]¶ An implicit feedback matrix factorization model. Uses a classic matrix factorization [1] approach, with latent vectors used to represent both users and items. Their dot product gives the predicted score for a useritem pair.
The latent representation is given by
spotlight.factorization.representations.BilinearNet
.The model is trained through negative sampling: for any known useritem pair, one or more items are randomly sampled to act as negatives (expressing a lack of preference by the user for the sampled item).
[1] Koren, Yehuda, Robert Bell, and Chris Volinsky. “Matrix factorization techniques for recommender systems.” Computer 42.8 (2009). Parameters:  loss (string, optional) – One of ‘pointwise’, ‘bpr’, ‘hinge’, or ‘adaptive hinge’,
corresponding to losses from
spotlight.losses
.  embedding_dim (int, optional) – Number of embedding dimensions to use for users and items.
 n_iter (int, optional) – Number of iterations to run.
 batch_size (int, optional) – Minibatch size.
 l2 (float, optional) – L2 loss penalty.
 learning_rate (float, optional) – Initial learning rate.
 optimizer_func (function, optional) – Function that takes in module parameters as the first argument and returns an instance of a PyTorch optimizer. Overrides l2 and learning rate if supplied. If no optimizer supplied, then use ADAM by default.
 use_cuda (boolean, optional) – Run the model on a GPU.
 representation (a representation module, optional) – If supplied, will override default settings and be used as the main network module in the model. Intended to be used as an escape hatch when you want to reuse the model’s training functions but want full freedom to specify your network topology.
 sparse (boolean, optional) – Use sparse gradients for embedding layers.
 random_state (instance of numpy.random.RandomState, optional) – Random state to use when fitting.
 num_negative_samples (int, optional) – Number of negative samples to generate for adaptive hinge loss.

fit
(interactions, verbose=False)[source]¶ Fit the model.
When called repeatedly, model fitting will resume from the point at which training stopped in the previous fit call.
Parameters:  interactions (
spotlight.interactions.Interactions
) – The input dataset.  verbose (bool) – Output additional information about current epoch and loss.
 interactions (

predict
(user_ids, item_ids=None)[source]¶ Make predictions: given a user id, compute the recommendation scores for items.
Parameters:  user_ids (int or array) – If int, will predict the recommendation scores for this user for all items in item_ids. If an array, will predict scores for all (user, item) pairs defined by user_ids and item_ids.
 item_ids (array, optional) – Array containing the item ids for which prediction scores are desired. If not supplied, predictions for all items will be computed.
Returns: predictions – Predicted scores for all items in item_ids.
Return type: np.array
 loss (string, optional) – One of ‘pointwise’, ‘bpr’, ‘hinge’, or ‘adaptive hinge’,
corresponding to losses from