bigframes.ml.impute.SimpleImputer#

class bigframes.ml.impute.SimpleImputer(strategy: Literal['mean', 'median', 'most_frequent'] = 'mean')[source]#

Univariate imputer for completing missing values with simple strategies.

Replace missing values using a descriptive statistic (e.g. mean, median, or most frequent) along each column.

Examples:

>>> import bigframes.pandas as bpd
>>> from bigframes.ml.impute import SimpleImputer
>>> X_train = bpd.DataFrame({"feat0": [7.0, 4.0, 10.0], "feat1": [2.0, None, 5.0], "feat2": [3.0, 6.0, 9.0]})
>>> imp_mean = SimpleImputer().fit(X_train)
>>> X_test = bpd.DataFrame({"feat0": [None, 4.0, 10.0], "feat1": [2.0, None, None], "feat2": [3.0, 6.0, 9.0]})
>>> imp_mean.transform(X_test)
   imputer_feat0  imputer_feat1  imputer_feat2
0            7.0            2.0            3.0
1            4.0            3.5            6.0
2           10.0            3.5            9.0

[3 rows x 3 columns]

Parameters:

strategy ({'mean', 'median', 'most_frequent'}, default='mean') The imputation strategy. mean: replace missing values using the mean along the axis. median:replace missing values using the median along the axis. most_frequent, replace missing using the most frequent value along the axis.

fit(X: DataFrame | Series | DataFrame | Series, y=None) → SimpleImputer[source]#

Fit the imputer on X.

Parameters:

• X (bigframes.dataframe.DataFrame or bigframes.series.Series or pandas.core.frame.DataFrame or pandas.core.series.Series) The Dataframe or Series with training data.

• y (default None) Ignored.

Returns:

Fitted scaler.

Return type:

SimpleImputer

transform(X: DataFrame | Series | DataFrame | Series) → DataFrame[source]#

Impute all missing values in X.

Parameters:

X (bigframes.dataframe.DataFrame or bigframes.series.Series or pandas.core.frame.DataFrame or pandas.core.series.Series) The DataFrame or Series to be transformed.

Returns:

Transformed result.

Return type:

bigframes.dataframe.DataFrame