Tabular Classification with Greedy Portfolio

The following example shows how to fit a sample classification model with AutoPyTorch using the greedy portfolio

import os
import tempfile as tmp
import warnings

os.environ['JOBLIB_TEMP_FOLDER'] = tmp.gettempdir()
os.environ['OMP_NUM_THREADS'] = '1'
os.environ['OPENBLAS_NUM_THREADS'] = '1'
os.environ['MKL_NUM_THREADS'] = '1'

warnings.simplefilter(action='ignore', category=UserWarning)
warnings.simplefilter(action='ignore', category=FutureWarning)

import sklearn.datasets
import sklearn.model_selection

from autoPyTorch.api.tabular_classification import TabularClassificationTask

Data Loading

X, y = sklearn.datasets.fetch_openml(data_id=40981, return_X_y=True, as_frame=True)
X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(
    X,
    y,
    random_state=42,
)

Build and fit a classifier

api = TabularClassificationTask(
    seed=42,
)

Search for an ensemble of machine learning algorithms

api.search(
    X_train=X_train,
    y_train=y_train,
    X_test=X_test.copy(),
    y_test=y_test.copy(),
    optimize_metric='accuracy',
    total_walltime_limit=300,
    func_eval_time_limit_secs=50,
    # Setting this option to "greedy"
    # will make smac run the configurations
    # present in 'autoPyTorch/configs/greedy_portfolio.json'
    portfolio_selection="greedy"
)

<autoPyTorch.api.tabular_classification.TabularClassificationTask object at 0x7f9aa6377ca0>

Print the final ensemble performance

y_pred = api.predict(X_test)
score = api.score(y_pred, y_test)
print(score)
# Print the final ensemble built by AutoPyTorch
print(api.show_models())

# Print statistics from search
print(api.sprint_statistics())

{'accuracy': 0.8786127167630058}
|    | Preprocessing                                                                                    | Estimator                                                          |   Weight |
|---:|:-------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------|---------:|
|  0 | SimpleImputer,Variance Threshold,MinorityCoalescer,OneHotEncoder,MinMaxScaler,PolynomialFeatures | no embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential    |     0.32 |
|  1 | SimpleImputer,Variance Threshold,NoCoalescer,OneHotEncoder,NoScaler,KernelPCA                    | no embedding,MLPBackbone,FullyConnectedHead,nn.Sequential          |     0.16 |
|  2 | SimpleImputer,Variance Threshold,NoCoalescer,OneHotEncoder,StandardScaler,NoFeaturePreprocessing | no embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential    |     0.14 |
|  3 | SimpleImputer,Variance Threshold,NoCoalescer,OneHotEncoder,MinMaxScaler,FastICA                  | no embedding,MLPBackbone,FullyConnectedHead,nn.Sequential          |     0.12 |
|  4 | SimpleImputer,Variance Threshold,NoCoalescer,OneHotEncoder,StandardScaler,NoFeaturePreprocessing | no embedding,ShapedResNetBackbone,FullyConnectedHead,nn.Sequential |     0.08 |
|  5 | None                                                                                             | CBLearner                                                          |     0.06 |
|  6 | None                                                                                             | ETLearner                                                          |     0.06 |
|  7 | SimpleImputer,Variance Threshold,NoCoalescer,OneHotEncoder,StandardScaler,NoFeaturePreprocessing | no embedding,ShapedResNetBackbone,FullyConnectedHead,nn.Sequential |     0.02 |
|  8 | SimpleImputer,Variance Threshold,NoCoalescer,OneHotEncoder,StandardScaler,NoFeaturePreprocessing | no embedding,ShapedResNetBackbone,FullyConnectedHead,nn.Sequential |     0.02 |
|  9 | None                                                                                             | RFLearner                                                          |     0.02 |
autoPyTorch results:
        Dataset name: 55bb3950-22f3-11ed-8835-b1fa420cf160
        Optimisation Metric: accuracy
        Best validation score: 0.8830409356725146
        Number of target algorithm runs: 25
        Number of successful target algorithm runs: 23
        Number of crashed target algorithm runs: 0
        Number of target algorithms that exceeded the time limit: 2
        Number of target algorithms that exceeded the memory limit: 0