jijzepttools.blackbox_optimization.surrogate_model.fmo_surrogate_model

jijzepttools.blackbox_optimization.surrogate_model.fmo_surrogate_model#

Classes#

FMParams

Parameters for Factorization Machine surrogate model.

FMSurrogateModel

Factorization Machine surrogate model implementation.

Module Contents#

class FMParams#

Bases: jijzepttools.blackbox_optimization.surrogate_model.interface.SurrogateModelParams

Parameters for Factorization Machine surrogate model.

latent_dim#

Dimension of latent factors in FM.

Type:

int, default=5

warm_start#

Whether to warm start the FM model. If True, the FM model will continue training from the previous weights. If False, the FM model will be re-initialized on each fit.

Type:

bool, default=True

verbose#

Whether to show training progress. Set to False to disable progress display.

Type:

bool, default=True

optimizer_config#

Configuration for the optimizer. Use AdamConfig for mini-batch SGD with Adam, or LBFGSConfig for full-batch quasi-Newton optimization.

Type:

AdamConfig | LBFGSConfig, default=LBFGSConfig()

latent_dim: int = 5#
warm_start: bool = True#
verbose: bool = True#
optimizer_config: jijzepttools.blackbox_optimization.fm_config.AdamConfig | jijzepttools.blackbox_optimization.fm_config.LBFGSConfig#
class FMSurrogateModel(n_features: int, params: FMParams | None = None, non_binary_indices: list[int] | None = None)#

Bases: jijzepttools.blackbox_optimization.surrogate_model.interface.SurrogateModel

Factorization Machine surrogate model implementation.

This model uses second-order feature interactions to approximate the objective function. The mathematical form is: f(x) = w0 + Σ wi*xi + ΣΣ <vi,vj>*xi*xj + Σ di*xi^2 (for non-binary)

Parameters:

  • n_features (int) – Number of input features.

  • params (FMParams, optional) – Parameters for the model. If None, default parameters are used.

  • non_binary_indices (list[int] | None, optional) – Indices of non-binary variables (integer/continuous). If provided, diagonal terms (x_i^2) will be added for these variables. For binary variables, x^2 = x, so diagonal terms are redundant with linear terms.

n_features#
non_binary_indices = None#
reset_state()#
fit(x: numpy.ndarray, y: numpy.ndarray) → None#

Fit the Factorization Machine to the data.

Parameters:

  • x (np.ndarray, shape (n_samples, n_features)) – Input features.

  • y (np.ndarray, shape (n_samples,)) – Target values.

predict(x: numpy.ndarray) → numpy.ndarray#

Predict target values using the fitted FM model.

surrogate_to_ommx_objective(decision_variables: list[ommx.v1.DecisionVariable])#

Create OMMX expression for the fitted FM model.

Parameters:

decision_variables (list[ommx.v1.DecisionVariable]) – Decision variables from OMMX instance.

Contents