FactorizationMachine¶

class FactorizationMachine¶

Bases: object

Factorization Machine model implemented in pure NumPy.

Methods

`__init__`	Initializes the Factorization Machine model.
`forward`	Computes the forward pass of the model.

Attributes

`bias`	Returns the bias term of the model.
`linear`	Returns the linear weights of the model.
`num_factors`	Returns the number of factors in the model.
`num_features`	Returns the number of input features in the model.
`quadratic`	Returns the quadratic weights of the model.

__call__(x: ndarray) → ndarray¶

Computes the forward pass of the model.

Parameters:: x (np.ndarray) – The input array of shape (n_samples, n_features).
Returns:: The output array of shape (n_samples,).
Return type:: np.ndarray

__init__(d: int, k: int, seed: int | SeedSequence | None = None) → None¶

Initializes the Factorization Machine model.

Parameters:

d (int) – The number of input features.
k (int) – The number of factors.
seed (int | np.random.SeedSequence | None, optional) – The seed for the random initialization. If None (default), a non-deterministic seed is used.

forward(x: ndarray) → ndarray¶

Computes the forward pass of the model.

Parameters:: x (np.ndarray) – The input array of shape (n_samples, n_features).
Returns:: The output array of shape (n_samples,).
Return type:: np.ndarray

__dict__ = mappingproxy({'__module__': 'amplify_bbopt.trainer', '__doc__': 'Factorization Machine model implemented in pure NumPy.', '__init__': <function FactorizationMachine.__init__>, 'num_factors': <property object>, 'num_features': <property object>, 'quadratic': <property object>, 'linear': <property object>, 'bias': <property object>, 'forward': <function FactorizationMachine.forward>, '__call__': <function FactorizationMachine.forward>, '__dict__': <attribute '__dict__' of 'FactorizationMachine' objects>, '__weakref__': <attribute '__weakref__' of 'FactorizationMachine' objects>, '__annotations__': {}})¶

property num_features: int¶: Returns the number of input features in the model.