TypeError：float() 参数必须是字符串或数字，而不是“SimpleImputer”

Question

all_data = pd.concat([train_data, test_data])
average = all_data.Age.median()
print('Average Age: {0}'.format(average))
train_data.fillna(value= {'Age' : average, inplace : True})
test_data.fillna(value= {'Age' : average, inplace : True})

错误：

TypeError                                 Traceback (most recent call
 last) <ipython-input-42-e465cf5beb42> in <module>()
       1 all_data = pd.concat([train_data, test_data])
 ----> 2 average = all_data.Age.median().values()
       3 print('Average Age: {0}'.format(average))
       4 train_data.fillna(value= {'Age' : average, inplace : True})
       5 test_data.fillna(value= {'Age' : average, inplace : True})
 
 3 frames /usr/local/lib/python3.7/dist-packages/pandas/core/nanops.py
 in nanmedian(values, axis, skipna, mask)
     611     if not is_float_dtype(values.dtype):
     612         try:
 --> 613             values = values.astype("f8")
     614         except ValueError as err:
     615             # e.g. "could not convert string to float: 'a'"
 
 TypeError: float() argument must be a string or a number, not
'SimpleImputer'

您好，我对这种情况感到困惑，即使使用 scikit-learn 也显示相同。请帮助我并提前感谢！

Answer 1

这应该可以解决问题：

median = all_data['Age'].median()
train_data['Age'] = train_data['Age'].fillna(value=median)
test_data['Age'] = test_data['Age'].fillna(value=median)