py3dinterpolations.modelling.estimator.Estimator

class py3dinterpolations.modelling.estimator.Estimator(griddata: GridData, params: dict, scoring: str = 'neg_mean_absolute_error', verbose: int = 3)

Bases: object

class for estimation of model parameters.

Runs a parameter estimation looking for the best scoring method. The scoring attribute can be any of the scoring methods supported by scikit-learn. See https://scikit-learn.org/stable/modules/model_evaluation.html#scoring-parameter

Get the best parameters with the best_params attribute. Get the best score with the best_score attribute. Get the cross validation results with the cv_results attribute.

Verbose level can be set from 0 to 3 to higher level of verbosity.

Note

At the moment only pykrige.Krige is supported.

Parameters

griddata (GridData) – griddata to interpolate
params (dict) – parameters to search
scoring (str, optional) – scoring method. Defaults to “neg_mean_absolute_error”.
verbose (int, optional) – verbosity level. Defaults to 3.

estimator

estimator object

Type: GridSearchCV

best_params

best parameters

Type: dict

best_score

best score

Type: float

cv_results

cross validation results, ready to be converted to pandas DataFrame

Type: dict

Examples

>>> # parameters of 3d kriging, both ordinary and universal
>>> params = {
>>>     "method": ["ordinary3d","universal3d"],
>>>     "variogram_model": ["linear", "power", "gaussian"],
>>>     "nlags": [2, 4, 6, 8, 10],
>>>     "weight": [True, False],
>>> }

Methods

__init__(griddata, params[, scoring, verbose])

Attributes

`best_params`
`best_score`
`cv_results`