def _fit_one_model(
self,
model: TimeSeriesForecastingModel,
X_split: pd.DataFrame,
results: pd.DataFrame,
only_model: bool = False,
) -> pd.DataFrame:
"""
Fits one model on a split and calculates its score and fit time
Parameters
----------
model: BaseEstimator, model to fit
X_split: pd.DataFrame, subset of training data to fit on
results: pd.DataFrame, results dataframe to add score results
only_model: bool, use ``only_model`` property to reuse the fitted features
Returns
-------
"""
start_time = time()
model_index = self._models_are_equal(model)
model.cache_features = True
model.fit(X_split, only_model=only_model)
scores = model.score(metrics=self.metrics)
results.loc[[model_index],
["Train score", "Test score"]] = scores.values
fit_time = time() - start_time
results.loc[model_index, "Fit time"] = fit_time
return results
def time_series_forecasting_model1_cache(features1, model1):
return TimeSeriesForecastingModel(
features=features1,
horizon=2,
model=model1,
cache_features=True,
)
def test_constructor(self, features1, model1):
horizon, cache_features = 2, True
time_series_forecasting_model = TimeSeriesForecastingModel(
features=features1,
horizon=horizon,
model=model1,
cache_features=cache_features,
)
assert time_series_forecasting_model.features == features1
assert time_series_forecasting_model.horizon == horizon
assert time_series_forecasting_model.model == model1
assert time_series_forecasting_model.cache_features == cache_features