sparse_ho.criterion.LogisticMulticlass

class sparse_ho.criterion.LogisticMulticlass(idx_train, idx_val, algo, idx_test=None)

Multiclass logistic loss.

Parameters

idx_train: ndarray

indices of the training set

idx_val: ndarray

indices of the validation set

algo: instance of ``sparse_ho.base.AlgoModel``

A model that follows the sparse_ho API.

idx_test: ndarray

indices of the test set

Attributes

dict_models: dict

dict with the models corresponding to each class.

__init__(idx_train, idx_val, algo, idx_test=None)

Methods

__init__(idx_train, idx_val, algo[, idx_test])
get_val(model, X, y, log_alpha, ...[, tol])	Get value of criterion.
get_val_grad(model, X, y, log_alpha, ...[, tol])	Get value and gradient of criterion.
grad_total_loss(all_betas, all_jacs, X, Y)	Compute the gradient of the multiclass logistic loss.
proj_hyperparam(model, X, y, log_alpha)	Project hyperparameter on admissible range of values