lenskit.implicit#

Bridges to recommendation models from implicit.

Note

This package is not included in the base lenskit distribution. Install lenskit-implicit from PyPI or Conda to get it.

Classes

ALS([config])

LensKit interface to implicit.cpu.als (or GPU version).

BPR([config])

LensKit interface to implicit.cpu.bpr (or GPU version).

BaseRec([config])

Base class for Implicit-backed recommenders.

ImplicitALSConfig(*[, weight])

ImplicitConfig(**extra_data)

class lenskit.implicit.BaseRec(config=None, **kwargs)#

Bases: Component[ItemList, …], Trainable

Base class for Implicit-backed recommenders.

Stability:
Caller (see Stability Levels).
Parameters:

  • config (ImplicitConfig)

  • kwargs (Any)

delegate: RecommenderBase#

The delegate algorithm from implicit.

matrix_: csr_matrix#

The user-item rating matrix from training.

users_: Vocabulary#

The user ID mapping from training.

items_: Vocabulary#

The item ID mapping from training.

train(data, options=TrainingOptions(retrain=True, device=None, rng=None))#

Train the model to learn its parameters from a training dataset.

Parameters:
class lenskit.implicit.ALS(config=None, **kwargs)#

Bases: BaseRec

LensKit interface to implicit.cpu.als (or GPU version).

Stability:
Caller (see Stability Levels).
Parameters:

  • config (ImplicitALSConfig)

  • kwargs (Any)

property weight#

Convert a string or number to a floating-point number, if possible.

class lenskit.implicit.BPR(config=None, **kwargs)#

Bases: BaseRec

LensKit interface to implicit.cpu.bpr (or GPU version).

Stability:
Caller (see Stability Levels).
Parameters:

  • config (ImplicitConfig)

  • kwargs (Any)