Start#

In this guide, we’ll show how you can use EvalML to automatically find the best pipeline for predicting whether or not a credit card transaction is fradulent. Along the way, we’ll highlight EvalML’s built-in tools and features for understanding and interacting with the search process.

[1]:
import evalml
from evalml import AutoMLSearch
from evalml.utils import infer_feature_types

First, we load in the features and outcomes we want to use to train our model.

[2]:
X, y = evalml.demos.load_fraud(n_rows=250)
             Number of Features
Boolean                       1
Categorical                   6
Numeric                       5

Number of training examples: 250
Targets
False    88.40%
True     11.60%
Name: fraud, dtype: object

First, we will clean the data. Since EvalML accepts a pandas input, it can run type inference on this data directly. Since we’d like to change the types inferred by EvalML, we can use the infer_feature_types utility method. Here’s what we’re going to do with the following dataset:

  • Reformat the expiration_date column so it reflects a more familiar date format.

  • Cast the lat and lng columns from float to str.

  • Use infer_feature_types to specify what types certain columns should be. For example, to avoid having the provider column be inferred as natural language text, we have specified it as a categorical column instead.

The infer_feature_types utility method takes a pandas or numpy input and converts it to a pandas dataframe with a Woodwork accessor, providing us with flexibility to cast the data as necessary.

[3]:
X.ww['expiration_date'] = X['expiration_date'].apply(lambda x: '20{}-01-{}'.format(x.split("/")[1], x.split("/")[0]))
X = infer_feature_types(X, feature_types= {'store_id': 'categorical',
                                           'expiration_date': 'datetime',
                                           'lat': 'categorical',
                                           'lng': 'categorical',
                                           'provider': 'categorical'})
X.ww
[3]:
Physical Type Logical Type Semantic Tag(s)
Column
card_id int64 Integer ['numeric']
store_id int64 Integer ['numeric']
datetime datetime64[ns] Datetime []
amount int64 Integer ['numeric']
currency string Unknown []
customer_present bool Boolean []
expiration_date datetime64[ns] Datetime []
provider category Categorical ['category']
lat float64 Double ['numeric']
lng float64 Double ['numeric']
region category Categorical ['category']
country category Categorical ['category']

In order to validate the results of the pipeline creation and optimization process, we will save some of our data as a holdout set.

[4]:
X_train, X_holdout, y_train, y_holdout = evalml.preprocessing.split_data(X, y, problem_type='binary', test_size=.2)

Note: To provide data to EvalML, it is recommended that you initialize a woodwork accessor so that you control how EvalML will treat each feature, such as as a numeric feature, a categorical feature, a text feature or other type of feature. Consult the the Woodwork project for help on how to do this. Here, split_data() returns dataframes with woodwork accessors.

EvalML has many options to configure the pipeline search. At the minimum, we need to define an objective function. For simplicity, we will use the F1 score in this example. However, the real power of EvalML is in using domain-specific objective functions or building your own.

Below EvalML utilizes Bayesian optimization (EvalML’s default optimizer) to search and find the best pipeline defined by the given objective.

EvalML provides a number of parameters to control the search process. max_batches is one of the parameters which controls the stopping criterion for the AutoML search. It indicates the maximum number of rounds of AutoML to evaluate, where each round may train and score a variable number of pipelines. In this example, max_batches is set to 1.

** Graphing methods, like AutoMLSearch, on Jupyter Notebook and Jupyter Lab require ipywidgets to be installed.

** If graphing on Jupyter Lab, jupyterlab-plotly required. To download this, make sure you have npm installed.

[5]:
automl = AutoMLSearch(X_train=X_train, y_train=y_train,
                      problem_type='binary', objective='f1',
                      max_batches=3, verbose=True)
Removing columns ['currency'] because they are of 'Unknown' type

When we call search(), the search for the best pipeline will begin. There is no need to wrangle with missing data or categorical variables as EvalML includes various preprocessing steps (like imputation, one-hot encoding, feature selection) to ensure you’re getting the best results. As long as your data is in a single table, EvalML can handle it. If not, you can reduce your data to a single table by utilizing Featuretools and its Entity Sets.

You can find more information on pipeline components and how to integrate your own custom pipelines into EvalML here.

[6]:
automl.search()
*****************************
* Beginning pipeline search *
*****************************

Optimizing for F1.
Greater score is better.

Using SequentialEngine to train and score pipelines.
Searching up to 3 batches for a total of None pipelines.
Allowed model families:

Evaluating Baseline Pipeline: Mode Baseline Binary Classification Pipeline
Mode Baseline Binary Classification Pipeline:
        Starting cross validation
        Finished cross validation - mean F1: 0.000

*****************************
* Evaluating Batch Number 1 *
*****************************

Logistic Regression Classifier w/ Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + One Hot Encoder + Oversampler + Standard Scaler:
        Starting cross validation
        Finished cross validation - mean F1: 0.227
Random Forest Classifier w/ Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + One Hot Encoder + Oversampler:
        Starting cross validation
        Finished cross validation - mean F1: 0.558

*****************************
* Evaluating Batch Number 2 *
*****************************

Logistic Regression Classifier w/ Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + One Hot Encoder + Oversampler + Standard Scaler + RF Classifier Select From Model:
        Starting cross validation
        Finished cross validation - mean F1: 0.276
Random Forest Classifier w/ Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + One Hot Encoder + Oversampler + RF Classifier Select From Model:
        Starting cross validation
        Finished cross validation - mean F1: 0.663

*****************************
* Evaluating Batch Number 3 *
*****************************

Decision Tree Classifier w/ Label Encoder + Select Columns By Type Transformer + Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + Oversampler + Select Columns Transformer + Select Columns Transformer + Label Encoder + Imputer + One Hot Encoder + Oversampler:
        Starting cross validation
        Finished cross validation - mean F1: 0.297
LightGBM Classifier w/ Label Encoder + Select Columns By Type Transformer + Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + Oversampler + Select Columns Transformer + Select Columns Transformer + Label Encoder + Imputer + One Hot Encoder + Oversampler:
        Starting cross validation
        Finished cross validation - mean F1: 0.589
Extra Trees Classifier w/ Label Encoder + Select Columns By Type Transformer + Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + Oversampler + Select Columns Transformer + Select Columns Transformer + Label Encoder + Imputer + One Hot Encoder + Oversampler:
        Starting cross validation
        Finished cross validation - mean F1: 0.365
Elastic Net Classifier w/ Label Encoder + Select Columns By Type Transformer + Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + Oversampler + Standard Scaler + Select Columns Transformer + Select Columns Transformer + Label Encoder + Imputer + One Hot Encoder + Oversampler + Standard Scaler:
        Starting cross validation
        Finished cross validation - mean F1: 0.235
CatBoost Classifier w/ Label Encoder + Select Columns By Type Transformer + Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + Oversampler + Select Columns Transformer + Select Columns Transformer + Label Encoder + Imputer + Oversampler:
        Starting cross validation
        Finished cross validation - mean F1: 0.683
XGBoost Classifier w/ Label Encoder + Select Columns By Type Transformer + Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + Oversampler + Select Columns Transformer + Select Columns Transformer + Label Encoder + Imputer + One Hot Encoder + Oversampler:
        Starting cross validation
        Finished cross validation - mean F1: 0.690

Search finished after 00:30
Best pipeline: XGBoost Classifier w/ Label Encoder + Select Columns By Type Transformer + Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + Oversampler + Select Columns Transformer + Select Columns Transformer + Label Encoder + Imputer + One Hot Encoder + Oversampler
Best pipeline F1: 0.689744

If you would like to suppress stdout output, set verbose=False. This is also the default behavior for AutoMLSearch if verbose is not specified.

[7]:
automl = AutoMLSearch(X_train=X_train, y_train=y_train,
                      problem_type='binary', objective='f1',
                      max_batches=3, verbose=False)
automl.search()

We also provide a standalone search method which does all of the above in a single line, and returns the AutoMLSearch instance and data check results. If there were data check errors, AutoML will not be run and no AutoMLSearch instance will be returned.

After the search is finished we can view all of the pipelines searched, ranked by score. Internally, EvalML performs cross validation to score the pipelines. If it notices a high variance across cross validation folds, it will warn you. EvalML also provides additional data checks to analyze your data to assist you in producing the best performing pipeline.

[8]:
automl.rankings
[8]:
id pipeline_name search_order mean_cv_score standard_deviation_cv_score validation_score percent_better_than_baseline high_variance_cv parameters
0 10 XGBoost Classifier w/ Label Encoder + Select C... 10 0.689744 0.165041 0.689744 68.974359 False {'Label Encoder': {'positive_label': None}, 'N...
1 9 CatBoost Classifier w/ Label Encoder + Select ... 9 0.682540 0.168919 0.682540 68.253968 False {'Label Encoder': {'positive_label': None}, 'N...
2 4 Random Forest Classifier w/ Label Encoder + Dr... 4 0.663337 0.263244 0.663337 66.333666 False {'Label Encoder': {'positive_label': None}, 'D...
3 6 LightGBM Classifier w/ Label Encoder + Select ... 6 0.588889 0.083887 0.588889 58.888889 False {'Label Encoder': {'positive_label': None}, 'N...
4 2 Random Forest Classifier w/ Label Encoder + Dr... 2 0.558405 0.182781 0.558405 55.840456 False {'Label Encoder': {'positive_label': None}, 'D...
5 7 Extra Trees Classifier w/ Label Encoder + Sele... 7 0.364957 0.085286 0.364957 36.495726 False {'Label Encoder': {'positive_label': None}, 'N...
6 5 Decision Tree Classifier w/ Label Encoder + Se... 5 0.296626 0.114974 0.296626 29.662639 False {'Label Encoder': {'positive_label': None}, 'N...
7 3 Logistic Regression Classifier w/ Label Encode... 3 0.275556 0.120247 0.275556 27.555556 False {'Label Encoder': {'positive_label': None}, 'D...
8 8 Elastic Net Classifier w/ Label Encoder + Sele... 8 0.235240 0.063793 0.235240 23.523954 False {'Label Encoder': {'positive_label': None}, 'N...
9 1 Logistic Regression Classifier w/ Label Encode... 1 0.227085 0.072648 0.227085 22.708518 False {'Label Encoder': {'positive_label': None}, 'D...
10 0 Mode Baseline Binary Classification Pipeline 0 0.000000 0.000000 0.000000 0.000000 False {'Label Encoder': {'positive_label': None}, 'B...

If we are interested in see more details about the pipeline, we can view a summary description using the id from the rankings table:

[9]:
automl.describe_pipeline(3)
**************************************************************************************************************************************************************************************************
* Logistic Regression Classifier w/ Label Encoder + Drop Columns Transformer + DateTime Featurizer + Imputer + One Hot Encoder + Oversampler + Standard Scaler + RF Classifier Select From Model *
**************************************************************************************************************************************************************************************************

Problem Type: binary
Model Family: Linear

Pipeline Steps
==============
1. Label Encoder
         * positive_label : None
2. Drop Columns Transformer
         * columns : ['currency']
3. DateTime Featurizer
         * features_to_extract : ['year', 'month', 'day_of_week', 'hour']
         * encode_as_categories : False
         * time_index : None
4. Imputer
         * categorical_impute_strategy : most_frequent
         * numeric_impute_strategy : mean
         * categorical_fill_value : None
         * numeric_fill_value : None
5. One Hot Encoder
         * top_n : 10
         * features_to_encode : None
         * categories : None
         * drop : if_binary
         * handle_unknown : ignore
         * handle_missing : error
6. Oversampler
         * sampling_ratio : 0.25
         * k_neighbors_default : 5
         * n_jobs : -1
         * sampling_ratio_dict : None
         * categorical_features : [3, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43]
         * k_neighbors : 5
7. Standard Scaler
8. RF Classifier Select From Model
         * number_features : None
         * n_estimators : 10
         * max_depth : None
         * percent_features : 0.5
         * threshold : median
         * n_jobs : -1
9. Logistic Regression Classifier
         * penalty : l2
         * C : 1.0
         * n_jobs : -1
         * multi_class : auto
         * solver : lbfgs

Training
========
Training for binary problems.
Objective to optimize binary classification pipeline thresholds for: <evalml.objectives.standard_metrics.F1 object at 0x7f51c4c36e50>
Total training time (including CV): 3.1 seconds

Cross Validation
----------------
               F1  MCC Binary  Log Loss Binary   Gini   AUC  Precision  Balanced Accuracy Binary  Accuracy Binary # Training # Validation
0           0.400       0.308            0.438  0.288 0.644      0.333                     0.682            0.821        133           67
1           0.267       0.175            0.609  0.208 0.604      0.286                     0.583            0.836        133           67
2           0.160       0.010            0.673 -0.017 0.492      0.111                     0.507            0.682        134           66
mean        0.276       0.164            0.573  0.160 0.580      0.243                     0.591            0.780          -            -
std         0.120       0.149            0.121  0.158 0.079      0.117                     0.088            0.085          -            -
coef of var 0.436       0.908            0.212  0.991 0.136      0.481                     0.149            0.109          -            -

We can also view the pipeline parameters directly:

[10]:
pipeline = automl.get_pipeline(3)
print(pipeline.parameters)
{'Label Encoder': {'positive_label': None}, 'Drop Columns Transformer': {'columns': ['currency']}, 'DateTime Featurizer': {'features_to_extract': ['year', 'month', 'day_of_week', 'hour'], 'encode_as_categories': False, 'time_index': None}, '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': 'if_binary', 'handle_unknown': 'ignore', 'handle_missing': 'error'}, 'Oversampler': {'sampling_ratio': 0.25, 'k_neighbors_default': 5, 'n_jobs': -1, 'sampling_ratio_dict': None, 'categorical_features': [3, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43], 'k_neighbors': 5}, 'RF Classifier Select From Model': {'number_features': None, 'n_estimators': 10, 'max_depth': None, 'percent_features': 0.5, 'threshold': 'median', 'n_jobs': -1}, 'Logistic Regression Classifier': {'penalty': 'l2', 'C': 1.0, 'n_jobs': -1, 'multi_class': 'auto', 'solver': 'lbfgs'}}

We can now select the best pipeline and score it on our holdout data:

[11]:
pipeline = automl.best_pipeline
pipeline.score(X_holdout, y_holdout, ["f1"])
[11]:
OrderedDict([('F1', 0.8)])

We can also visualize the structure of the components contained by the pipeline:

[12]:
pipeline.graph()
[12]:
_images/start_26_0.svg