eland.ml.ImportedMLModel.predict

ImportedMLModel.predict(X: Union[List[float], List[List[float]]]) → numpy.ndarray

Make a prediction using a trained model stored in Elasticsearch.

Parameters for this method are not yet fully compatible with standard sklearn.predict.

Parameters

X: list or list of lists of type float

Input feature vector - TODO support DataFrame and other formats

Returns

y: np.ndarray of dtype float for regressors or int for classifiers

Examples

>>> from sklearn import datasets
>>> from xgboost import XGBRegressor
>>> from eland.ml import ImportedMLModel

>>> # Train model
>>> training_data = datasets.make_classification(n_features=6, random_state=0)
>>> test_data = [[-1, -2, -3, -4, -5, -6], [10, 20, 30, 40, 50, 60]]
>>> regressor = XGBRegressor(objective='reg:squarederror')
>>> regressor = regressor.fit(training_data[0], training_data[1])

>>> # Get some test results
>>> regressor.predict(np.array(test_data))
array([0.06062475, 0.9990102 ], dtype=float32)

>>> # Serialise the model to Elasticsearch
>>> feature_names = ["f0", "f1", "f2", "f3", "f4", "f5"]
>>> model_id = "test_xgb_regressor"
>>> es_model = ImportedMLModel('localhost', model_id, regressor, feature_names, es_if_exists='replace')

>>> # Get some test results from Elasticsearch model
>>> es_model.predict(test_data)
array([0.0606248 , 0.99901026], dtype=float32)

>>> # Delete model from Elasticsearch
>>> es_model.delete_model()

eland.ml.ImportedMLModel eland.ml.MLModel.exists_model