utils#

Utility methods for EvalML components.

Module Contents#

Classes Summary#

`WrappedSKClassifier`	Scikit-learn classifier wrapper class.
`WrappedSKRegressor`	Scikit-learn regressor wrapper class.

Functions#

`all_components`	Get all available components.
`allowed_model_families`	List the model types allowed for a particular problem type.
`estimator_unable_to_handle_nans`	If True, provided estimator class is unable to handle NaN values as an input.
`generate_component_code`	Creates and returns a string that contains the Python imports and code required for running the EvalML component.
`get_estimators`	Returns the estimators allowed for a particular problem type.
`get_prediction_intevals_for_tree_regressors`	Find the prediction intervals for tree-based regressors.
`handle_component_class`	Standardizes input from a string name to a ComponentBase subclass if necessary.
`handle_float_categories_for_catboost`	Updates input data to be compatible with CatBoost estimators.
`make_balancing_dictionary`	Makes dictionary for oversampler components. Find ratio of each class to the majority. If the ratio is smaller than the sampling_ratio, we want to oversample, otherwise, we don't want to sample at all, and we leave the data as is.
`scikit_learn_wrapped_estimator`	Wraps an EvalML object as a scikit-learn estimator.

Contents#

evalml.pipelines.components.utils.all_components()[source]#: Get all available components.

evalml.pipelines.components.utils.allowed_model_families(problem_type)[source]#

List the model types allowed for a particular problem type.

Parameters: problem_type (ProblemTypes or str) – ProblemTypes enum or string.
Returns: A list of model families.
Return type: list[ModelFamily]

evalml.pipelines.components.utils.estimator_unable_to_handle_nans(estimator_class)[source]#

If True, provided estimator class is unable to handle NaN values as an input.

Parameters: estimator_class (Estimator) – Estimator class
Raises: ValueError – If estimator is not a valid estimator class.
Returns: True if estimator class is unable to process NaN values, False otherwise.
Return type: bool

evalml.pipelines.components.utils.generate_component_code(element)[source]#

Creates and returns a string that contains the Python imports and code required for running the EvalML component.

Parameters: element (component instance) – The instance of the component to generate string Python code for.
Returns: String representation of Python code that can be run separately in order to recreate the component instance. Does not include code for custom component implementation.
Raises: ValueError – If the input element is not a component instance.

Examples

>>> from evalml.pipelines.components.estimators.regressors.decision_tree_regressor import DecisionTreeRegressor
>>> assert generate_component_code(DecisionTreeRegressor()) == "from evalml.pipelines.components.estimators.regressors.decision_tree_regressor import DecisionTreeRegressor\n\ndecisionTreeRegressor = DecisionTreeRegressor(**{'criterion': 'squared_error', 'max_features': 'auto', 'max_depth': 6, 'min_samples_split': 2, 'min_weight_fraction_leaf': 0.0})"
...
>>> from evalml.pipelines.components.transformers.imputers.simple_imputer import SimpleImputer
>>> assert generate_component_code(SimpleImputer()) == "from evalml.pipelines.components.transformers.imputers.simple_imputer import SimpleImputer\n\nsimpleImputer = SimpleImputer(**{'impute_strategy': 'most_frequent', 'fill_value': None})"

evalml.pipelines.components.utils.get_estimators(problem_type, model_families=None, excluded_model_families=None)[source]#

Returns the estimators allowed for a particular problem type.

Can also optionally filter by a list of model types.

Parameters

problem_type (ProblemTypes or str) – Problem type to filter for.
model_families (list[ModelFamily] or list[str]) – Model families to filter for.
excluded_model_families (list[ModelFamily]) – A list of model families to exclude from the results.

Returns

A list of estimator subclasses.

Return type

list[class]

Raises

TypeError – If the model_families parameter is not a list.
RuntimeError – If a model family is not valid for the problem type.

evalml.pipelines.components.utils.get_prediction_intevals_for_tree_regressors(X: pandas.DataFrame, predictions: pandas.Series, coverage: List[float], estimators: List[evalml.pipelines.components.estimators.estimator.Estimator]) → Dict[str, pandas.Series][source]#

Find the prediction intervals for tree-based regressors.

Parameters

X (pd.DataFrame) – Data of shape [n_samples, n_features].
predictions (pd.Series) – Predictions from the regressor.
coverage (list[float]) – A list of floats between the values 0 and 1 that the upper and lower bounds of the prediction interval should be calculated for.
estimators (list) – Collection of fitted sub-estimators.

Returns

Prediction intervals, keys are in the format {coverage}_lower or {coverage}_upper.

Return type

dict

evalml.pipelines.components.utils.handle_component_class(component_class)[source]#

Standardizes input from a string name to a ComponentBase subclass if necessary.

If a str is provided, will attempt to look up a ComponentBase class by that name and return a new instance. Otherwise if a ComponentBase subclass or Component instance is provided, will return that without modification.

Parameters

component_class (str, ComponentBase) – Input to be standardized.

Returns

ComponentBase

Raises

ValueError – If input is not a valid component class.
MissingComponentError – If the component cannot be found.

Examples

>>> from evalml.pipelines.components.estimators.regressors.decision_tree_regressor import DecisionTreeRegressor
>>> handle_component_class(DecisionTreeRegressor)
<class 'evalml.pipelines.components.estimators.regressors.decision_tree_regressor.DecisionTreeRegressor'>
>>> handle_component_class("Random Forest Regressor")
<class 'evalml.pipelines.components.estimators.regressors.rf_regressor.RandomForestRegressor'>

evalml.pipelines.components.utils.handle_float_categories_for_catboost(X)[source]#

Updates input data to be compatible with CatBoost estimators.

CatBoost cannot handle data in X that is the Categorical Woodwork logical type with floating point categories. This utility determines if the floating point categories can be converted to integers without truncating any data, and if they can be, converts them to int64 categories. Will not attempt to use values that are truly floating points.

Parameters

X (pd.DataFrame) – Input data to CatBoost that has Woodwork initialized

Returns

Input data with exact same Woodwork typing info as the original but with any float categories: converted to be int64 when possible.

Return type

DataFrame

Raises

ValueError – if the numeric categories are actual floats that cannot be converted to integers without truncating data

evalml.pipelines.components.utils.make_balancing_dictionary(y, sampling_ratio)[source]#

Makes dictionary for oversampler components. Find ratio of each class to the majority. If the ratio is smaller than the sampling_ratio, we want to oversample, otherwise, we don’t want to sample at all, and we leave the data as is.

Parameters

y (pd.Series) – Target data.
sampling_ratio (float) – The balanced ratio we want the samples to meet.

Returns

Dictionary where keys are the classes, and the corresponding values are the counts of samples for each class that will satisfy sampling_ratio.

Return type

dict

Raises

ValueError – If sampling ratio is not in the range (0, 1] or the target is empty.

Examples

>>> import pandas as pd
>>> y = pd.Series([1] * 4 + [2] * 8 + [3])
>>> assert make_balancing_dictionary(y, 0.5) == {2: 8, 1: 4, 3: 4}
>>> assert make_balancing_dictionary(y, 0.9) == {2: 8, 1: 7, 3: 7}
>>> assert make_balancing_dictionary(y, 0.1) == {2: 8, 1: 4, 3: 1}

evalml.pipelines.components.utils.scikit_learn_wrapped_estimator(evalml_obj)[source]#: Wraps an EvalML object as a scikit-learn estimator.

class evalml.pipelines.components.utils.WrappedSKClassifier(pipeline)[source]#

Scikit-learn classifier wrapper class.

Methods

`fit`	Fits component to data.
`get_params`	Get parameters for this estimator.
`predict`	Make predictions using selected features.
`predict_proba`	Make probability estimates for labels.
`score`	Return the mean accuracy on the given test data and labels.
`set_params`	Set the parameters of this estimator.

fit(self, X, y)[source]#

Fits component to data.

Parameters

X (pd.DataFrame or np.ndarray) – The input training data of shape [n_samples, n_features].
y (pd.Series, optional) – The target training data of length [n_samples].

Returns

self

get_params(self, deep=True)#

Get parameters for this estimator.

Parameters: deep (bool, default=True) – If True, will return the parameters for this estimator and contained subobjects that are estimators.
Returns: params – Parameter names mapped to their values.
Return type: dict

predict(self, X)[source]#

Make predictions using selected features.

Parameters: X (pd.DataFrame) – Features
Returns: Predicted values.
Return type: np.ndarray

predict_proba(self, X)[source]#

Make probability estimates for labels.

Parameters: X (pd.DataFrame) – Features.
Returns: Probability estimates.
Return type: np.ndarray

score(self, X, y, sample_weight=None)#

Return the mean accuracy on the given test data and labels.

In multi-label classification, this is the subset accuracy which is a harsh metric since you require for each sample that each label set be correctly predicted.

Parameters

X (array-like of shape (n_samples, n_features)) – Test samples.
y (array-like of shape (n_samples,) or (n_samples, n_outputs)) – True labels for X.
sample_weight (array-like of shape (n_samples,), default=None) – Sample weights.

Returns

score – Mean accuracy of self.predict(X) w.r.t. y.

Return type

float

set_params(self, **params)#

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects (such as Pipeline). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

Parameters: **params (dict) – Estimator parameters.
Returns: self – Estimator instance.
Return type: estimator instance

class evalml.pipelines.components.utils.WrappedSKRegressor(pipeline)[source]#

Scikit-learn regressor wrapper class.

Methods

`fit`	Fits component to data.
`get_params`	Get parameters for this estimator.
`predict`	Make predictions using selected features.
`score`	Return the coefficient of determination of the prediction.
`set_params`	Set the parameters of this estimator.

fit(self, X, y)[source]#

Fits component to data.

Parameters

X (pd.DataFrame or np.ndarray) – the input training data of shape [n_samples, n_features]
y (pd.Series, optional) – the target training data of length [n_samples]

Returns

self

get_params(self, deep=True)#

Get parameters for this estimator.

Parameters: deep (bool, default=True) – If True, will return the parameters for this estimator and contained subobjects that are estimators.
Returns: params – Parameter names mapped to their values.
Return type: dict

predict(self, X)[source]#

Make predictions using selected features.

Parameters: X (pd.DataFrame) – Features.
Returns: Predicted values.
Return type: np.ndarray

score(self, X, y, sample_weight=None)#

Return the coefficient of determination of the prediction.

The coefficient of determination \(R^2\) is defined as \((1 - \frac{u}{v})\), where \(u\) is the residual sum of squares ((y_true - y_pred)** 2).sum() and \(v\) is the total sum of squares ((y_true - y_true.mean()) ** 2).sum(). The best possible score is 1.0 and it can be negative (because the model can be arbitrarily worse). A constant model that always predicts the expected value of y, disregarding the input features, would get a \(R^2\) score of 0.0.

Parameters

X (array-like of shape (n_samples, n_features)) – Test samples. For some estimators this may be a precomputed kernel matrix or a list of generic objects instead with shape (n_samples, n_samples_fitted), where n_samples_fitted is the number of samples used in the fitting for the estimator.
y (array-like of shape (n_samples,) or (n_samples, n_outputs)) – True values for X.
sample_weight (array-like of shape (n_samples,), default=None) – Sample weights.

Returns

score – \(R^2\) of self.predict(X) w.r.t. y.

Return type

float

Notes

The \(R^2\) score used when calling score on a regressor uses multioutput='uniform_average' from version 0.23 to keep consistent with default value of r2_score(). This influences the score method of all the multioutput regressors (except for MultiOutputRegressor).

set_params(self, **params)#

Set the parameters of this estimator.

Parameters: **params (dict) – Estimator parameters.
Returns: self – Estimator instance.
Return type: estimator instance