ColumnTransformer失败，管道中有CountVectorizer/HashingVectorizer(多个文本功能)

原学程将引见ColumnTransformer掉败，管讲中有CountVectorizer/HashingVectorizer(多个文原功效)的处置办法，这篇学程是从其余处所瞅到的，而后减了1些海外法式员的疑问与解问，愿望能对于您有所赞助，佳了，上面开端进修吧。

成绩描写

相似于此成绩(ColumnTransformer fails with CountVectorizer in a pipeline)，我愿望应用管讲中的ColumnTransformer对于具备文原功效的列运用CountVectorizer/HashingVectorizer。但是我没有是只要1个文字功效，而是多个。假如我传播了1个功效(而没有是像另外一个成绩的处理计划中修议的这样作为列表)，它任务患上很佳，我怎样为多个功效传播它？

numeric_features = ['x0', 'x一', 'y0', 'y一']
categorical_features = []
text_features = ['text_feature', 'another_text_feature']

numeric_transformer = Pipeline(steps=[('scaler', StandardScaler())])
categorical_transformer = Pipeline(steps=[('encoder', OneHotEncoder())])
text_transformer = Pipeline(steps=[('hashing', HashingVectorizer())])

preprocessor = ColumnTransformer(transformers=[
 ('numeric', numeric_transformer, numeric_features), 
 ('categorical', categorical_transformer, categorical_features),
 ('text', text_transformer, text_features)
])
 
steps = [('preprocessor', preprocessor),
('clf', SGDClassifier())]
 
pipeline = Pipeline(steps=steps)
 
pipeline.fit(X_train, y_train)

推举谜底

只需为每一个文原功效应用零丁的转换器。

preprocessor = ColumnTransformer(transformers=[
 ('numeric', numeric_transformer, numeric_features), 
 ('categorical', categorical_transformer, categorical_features),
 ('text', text_transformer, 'text_feature'),
 ('more_text', text_transformer, 'another_text_feature'),
])

(转换器在拆卸进程中被克隆，是以叨教有二个零丁的text_transformer正本，1切皆很佳。假如您担忧像如许指定雷同的转换器二次，您一直不妨在指定ColumnTransformer之前脚动复制/克隆它。)

佳了闭于ColumnTransformer掉败，管讲中有CountVectorizer/HashingVectorizer(多个文原功效)的学程便到这里便停止了，愿望趣模板源码网找到的这篇技巧文章能赞助到年夜野，更多技巧学程不妨在站内搜刮。