关于value_count

1 train_data = load_titanic_data("train.csv")
2 train_data["Pclass"].value_counts()

3 491 
1 216 
2 184 
Name: Pclass, dtype: int64

 1 from sklearn.pipeline import Pipeline
 2 from sklearn.preprocessing import Imputer
 3 
 4 num_pipeline = Pipeline([
 5         ("select_numeric", DataFrameSelector(["Age", "SibSp", "Parch", "Fare"])),
 6         ("imputer", Imputer(strategy="median")),
 7     ])
 8 
 9 class MostFrequentImputer(BaseEstimator, TransformerMixin):
10     def fit(self, X, y=None):
11         self.most_frequent_ = pd.Series([X[c].value_counts().index[0] for c in X],
12                                         index=X.columns)
13         return self
14     def transform(self, X, y=None):
15         return X.fillna(self.most_frequent_)
16 
17 from future_encoders import OneHotEncoder
18 cat_pipeline = Pipeline([
19         ("select_cat", DataFrameSelector(["Pclass", "Sex", "Embarked"])),
20         ("imputer", MostFrequentImputer()),
21         ("cat_encoder", OneHotEncoder(sparse=False)),
22     ])
23 
24 cat_pipeline.fit_transform(train_data)