Dask Dataframe 一热编码

Question

如何将 ddf 标签转换为 Dask 数据帧中的一种热编码？ 标签是从 1 到 17 的整数

这里是我的代码：

df = dask.dataframe.read_csv('nonTorDataset2.csv', dtype=dicType)

train, test = dask_ml.model_selection.train_test_split(df,test_size=0.2,shuffle=True,blockwise=True)
y_test = train['class_id']
y_test=y_test.astype('category')
y_test=y_test.categorize()
encoder = OneHotEncoder(sparse=False)
encoder = encoder.fit(y_test)
print(encoder.fit(y_test).compute())

              

                            

Answer 1

您可以使用DummyEncoder。

from dask_ml.preprocessing import DummyEncoder

# First categorize your columns.  
ddf = ddf.categorize(["class_id"]) 

# Then generate an encoder. 
encoder = DummyEncoder() 

# Fit the encoder 
ddf_enc = encoder.fit_transform(ddf)  

# Take a look at the data  
ddf_enc.head()