skrub.DataOp.skb.make_learner#

DataOp.skb.make_learner(*, fitted=False, keep_subsampling=False, choose='default')[source]#

Get a skrub learner for this DataOp.

Returns a SkrubLearner with a fit() method so it can be fit to some training data and then apply it to unseen data by calling transform() or predict(). Unlike scikit-learn estimators, skrub learners accept a dictionary of inputs rather than X and y arguments.

Warning

If the DataOp contains choices (e.g. choose_from(...)), by default this learner uses the default value of each choice. See the choose parameter for other options (random or from an Optuna trial). To actually pick the best value with hyperparameter tuning, use DataOp.skb.make_randomized_search() DataOp.skb.make_grid_search() instead, or an Optuna Study as shown in this example.

Parameters:

fittedbool (default=False)

If true, the returned learner is fitted to the data provided when initializing variables in skrub.var("name", value=...) and skrub.X(value).

keep_subsamplingbool (default=False)

If True, and if subsampling has been configured (see DataOp.skb.subsample()), fit on a subsample of the data. By default subsampling is not applied and all the data is used. This is only applied for fitting the estimator when fitted=True, subsequent use of the estimator is not affected by subsampling. Therefore it is an error to pass keep_subsampling=True and fitted=False (because keep_subsampling=True would have no effect).

choose‘default’, ‘random’, ‘random([seed])’ or optuna.Trial instance

How to resolve choices contained in the data_op. The different options are:

‘default’: the corresponding parameters of the SkrubLearner are not set; the default values of the choices are used.
‘random’: a random value is picked according to the distribution of each choice. The form ‘random([seed])’ is also accepted to set the random seed: for example ‘random(0)’ sets it to 0. ‘random()’ is the same as ‘random’.
an instance of numpy.random.RandomState. Same as ‘random’, but the provided RandomState is used to sample values.
an instance of optuna.Trial or optuna.FrozenTrial. It is used to suggest values for the choices.

Note that none of these options picks the best choice value according to an evaluation criterion, as this function creates a single learner. These options can be combined with external logic to evaluate and select the resulting learners, or one of DataOp.skb.make_grid_search(), DataOp.skb.make_randomized_search(), optuna.Study.optimize (as shown in this example) can be used to automatically select the best hyperparameters.

Returns:

learner: A skrub learner with an interface similar to scikit-learn’s, except that its methods accept a dictionary of named inputs rather than X and y arguments.

Examples

>>> import skrub
>>> from sklearn.dummy import DummyClassifier
>>> orders_df = skrub.datasets.toy_orders(split="train").orders
>>> orders = skrub.var('orders', orders_df)
>>> X = orders.drop(columns='delayed', errors='ignore').skb.mark_as_X()
>>> y = orders['delayed'].skb.mark_as_y()
>>> pred = X.skb.apply(skrub.TableVectorizer()).skb.apply(
...     DummyClassifier(), y=y
... )
>>> pred
<Apply DummyClassifier>
Result:
―――――――
0    False
1    False
2    False
3    False
Name: delayed, dtype: bool
>>> learner = pred.skb.make_learner(fitted=True)
>>> new_orders_df = skrub.datasets.toy_orders(split='test').X
>>> new_orders_df
   ID product  quantity        date
4   5     cup         5  2020-04-11
5   6    fork         2  2020-04-12
>>> learner.predict({'orders': new_orders_df})
array([False, False])

Note that the 'orders' key in the dictionary passed to predict corresponds to the name 'orders' in skrub.var('orders', orders_df) above.

The choose parameter allows us to control how choices contained in the DataOp should be handled. The default is to use the default value of each choice.

>>> def mult(x, factor):
...     return x * factor
>>> out = skrub.var("x").skb.apply_func(
...     mult, skrub.choose_int(-10, 10, default=2)
... )
>>> out.skb.make_learner().fit_transform({'x': 1})
2

The ‘random’ option samples new choice outcomes for each created learner:

>>> out.skb.make_learner(choose='random').fit_transform({'x': 1})
np.int64(3)
>>> out.skb.make_learner(choose='random').fit_transform({'x': 1})
np.int64(-5)

If an optuna.Trial instance is passed instead, the choice outcomes are obtained by calling the trial’s suggest_int, suggest_float or suggest_categorical methods. This allows easily selecting hyperparameters with optuna.

Please see the examples gallery for full information about DataOps and the learners they generate.

skrub.DataOp.skb.make_learner#

This Page