EvalML pipelines represent a sequence of operations to be applied to data, where each operation is either a data transformation or an ML modeling algorithm.

A pipeline class holds a combination of one or more components, which will be applied to new input data in sequence.

Each component and pipeline class supports a set of parameters which configure its behavior. The AutoML search process seeks to find the combination of pipeline structure and pipeline parameters which perform the best on the data.

Class Definition

Pipeline definitions must inherit from the proper pipeline base class, RegressionPipeline, BinaryClassificationPipeline or MulticlassClassificationPipeline. They must also include a component_graph list as a class variable containing the sequence of components to be fit and evaluated. The component_graph list is used to determine the ordered list of components that should be instantiated when a pipeline instance is created. Each component in component_graph can be provided as a reference to the component class for custom components, and as either a string name or as a reference to the component class for components defined in EvalML.

from evalml.pipelines import MulticlassClassificationPipeline

class CustomMulticlassClassificationPipeline(MulticlassClassificationPipeline):
    component_graph = ['Imputer', 'Random Forest Classifier']

If you’re using your own custom components you can refer to them like so:

from evalml.pipelines.components import Transformer

class NewTransformer(Transformer):
    name = 'New Transformer'
    hyperparameter_ranges = {
        "parameter_1":['a', 'b', 'c']

    def __init__(self, parameter_1, random_state):
        transformer = ThirdPartyTransformer(parameter_1)
        parameters = {"parameter_1": parameter_1}

class CustomComponentMulticlassClassificationPipeline(MulticlassClassificationPipeline):
    component_graph = [NewTransformer, 'Random Forest Classifier']

Pipeline Usage

All pipelines define the following methods:

  • fit fits each component on the provided training data, in order.

  • predict computes the predictions of the component graph on the provided data.

  • score computes the value of an objective on the provided data.

from evalml.demos import load_wine
X, y = load_wine()

pipeline = CustomMulticlassClassificationPipeline({})
pipeline.fit(X, y)
print(pipeline.score(X, y, objectives=['log loss multiclass']))
0      class_0
1      class_0
2      class_0
3      class_0
4      class_0
173    class_2
174    class_2
175    class_2
176    class_2
177    class_2
Length: 178, dtype: object
OrderedDict([('Log Loss Multiclass', 0.04132737017536148)])

Custom Name

By default, a pipeline class’s name property is the result of adding spaces between each Pascal case capitalization in the class name. E.g. LogisticRegressionPipeline.name will return ‘Logistic Regression Pipeline’. Therefore, we suggest custom pipelines use Pascal case for their class names.

If you’d like to override the pipeline classes name attribute so it isn’t derived from the class name, you can set the custom_name attribute, like so:

from evalml.pipelines import MulticlassClassificationPipeline

class CustomPipeline(MulticlassClassificationPipeline):
    component_graph = ['Imputer', 'One Hot Encoder', 'Logistic Regression Classifier']
    custom_name = 'A custom pipeline name'

A custom pipeline name

Override Component Hyperparameter Ranges

To specify custom hyperparameter ranges, set the custom_hyperparameters property to be a dictionary where each key-value pair consists of a parameter name and range. AutoML will use this dictionary to override the hyperparameter ranges collected from each component in the component graph.

If the hyperparameter ranges are categorical values, they can be passed in as lists or as skopt.space.Categorical values.

from skopt.space import Categorical
class CustomPipeline(MulticlassClassificationPipeline):
    component_graph = ['Imputer', 'One Hot Encoder', 'Standard Scaler', 'Logistic Regression Classifier']

print("Without custom hyperparameters:")

class CustomPipeline(MulticlassClassificationPipeline):
        component_graph = ['Imputer', 'One Hot Encoder', 'Standard Scaler', 'Logistic Regression Classifier']
        custom_hyperparameters = {
        'Simple Imputer' : {
            'impute_strategy': Categorical(['most_frequent']),
            # Can also pass in a list, like below
            'another_hyperparameter': ['value']

print("With custom hyperparameters:")
Without custom hyperparameters:
{'Imputer': {'categorical_impute_strategy': ['most_frequent'], 'numeric_impute_strategy': ['mean', 'median', 'most_frequent']}, 'One Hot Encoder': {}, 'Standard Scaler': {}, 'Logistic Regression Classifier': {'penalty': ['l2'], 'C': Real(low=0.01, high=10, prior='uniform', transform='identity')}}

With custom hyperparameters:
{'Imputer': {'categorical_impute_strategy': ['most_frequent'], 'numeric_impute_strategy': ['mean', 'median', 'most_frequent']}, 'One Hot Encoder': {}, 'Standard Scaler': {}, 'Logistic Regression Classifier': {'penalty': ['l2'], 'C': Real(low=0.01, high=10, prior='uniform', transform='identity')}}

To initialize our new custom pipeline class, we must pass in a parameters argument. If we want to use the defaults for each component, we can simply pass in an empty dictionary.

CustomPipeline(parameters={'Imputer':{'categorical_impute_strategy': 'most_frequent', 'numeric_impute_strategy': 'mean', 'categorical_fill_value': None, 'numeric_fill_value': None}, 'One Hot Encoder':{'top_n': 10, 'features_to_encode': None, 'categories': None, 'drop': None, 'handle_unknown': 'ignore', 'handle_missing': 'error'}, 'Logistic Regression Classifier':{'penalty': 'l2', 'C': 1.0, 'n_jobs': -1, 'multi_class': 'auto', 'solver': 'lbfgs'},})

Pipeline Parameters

You can also pass in custom parameters, which will then be used when instantiating each component in component_graph. The parameters dictionary needs to be in the format of a two-layered dictionary where the key-value pairs are the component name and corresponding component parameters dictionary. The component parameters dictionary consists of (parameter name, parameter values) key-value pairs.

An example will be shown below. The API reference for component parameters can also be found here.

parameters = {
        'Imputer': {
            "categorical_impute_strategy": "most_frequent",
            "numeric_impute_strategy": "median"
        'Logistic Regression Classifier': {
            'penalty': 'l2',
            'C': 1.0,

cp = CustomPipeline(parameters=parameters, random_state=5)

Pipeline Description

You can call .graph() to see each component and its parameters. Each component takes in data and feeds it to the next.


You can see a textual representation of the pipeline by calling .describe():

* Custom Pipeline *

Problem Type: multiclass
Model Family: Linear

Pipeline Steps
1. Imputer
         * categorical_impute_strategy : most_frequent
         * numeric_impute_strategy : median
         * categorical_fill_value : None
         * numeric_fill_value : None
2. One Hot Encoder
         * top_n : 10
         * features_to_encode : None
         * categories : None
         * drop : None
         * handle_unknown : ignore
         * handle_missing : error
3. Standard Scaler
4. Logistic Regression Classifier
         * penalty : l2
         * C : 1.0
         * n_jobs : -1
         * multi_class : auto
         * solver : lbfgs

Component Graph

You can use the pipeline’s component_graph attribute to access a component at a specific index:

first_component = cp.component_graph[0]
print (first_component.name)

Alternatively, you can use pipeline.get_component(name) and provide the component name instead (API reference here):

Imputer(categorical_impute_strategy='most_frequent', numeric_impute_strategy='median', categorical_fill_value=None, numeric_fill_value=None)

Pipeline Estimator

EvalML enforces that the last component of a pipeline is an estimator. You can access this estimator directly by using either pipeline.component_graph[-1] or pipeline.estimator.

LogisticRegressionClassifier(penalty='l2', C=1.0, n_jobs=-1, multi_class='auto', solver='lbfgs')
LogisticRegressionClassifier(penalty='l2', C=1.0, n_jobs=-1, multi_class='auto', solver='lbfgs')

Input Feature Names

After a pipeline is fitted, you can access a pipeline’s input_feature_names attribute to obtain a dictionary containing a list of feature names passed to each component of the pipeline. This could be especially useful for debugging where a feature might have been dropped or detecting unexpected behavior.

{'Imputer': ['alcohol',
 'Random Forest Classifier': ['alcohol',