Python替换缺失值

Question

我正在尝试替换数据框中特定列中的缺失值，但遇到了一些问题。 试过了：

from sklearn.impute import SimpleImputer
fill_0_with_mean = SimpleImputer(missing_values=0, strategy='mean')
X_train['Age'] = fill_0_with_mean.fit_transform(X_train['Age'])

和

X_train[:,15] = fill_0_with_mean.fit_transform(X_train[:,15])

和

X_train[:,15:16] = fill_0_with_mean.fit_transform(X_train[:,15:16])

和

X_train['Age'] = fill_0_with_mean.fit_transform(X_train['Age'].values)

和

X_train[:,15:16] = fill_0_with_mean.fit_transform(X_train[:,15:16].values)

但我总是遇到错误 ValueError: Input contains NaN, infinity or a value too large for dtype('float64'). or IndexError: only integers, slices (:), ellipsis (...), numpy.newaxis (None) 和整数或布尔数组是有效的索引

我的数据中有零值和缺失 (NaN) 值。 imputer只能做两者之一吗？我该怎么做呢？ 我也尝试将我的年龄列转换为整数

X_train['Age'] = X_train['Age'].as_type('int32')

但这只会给我其他错误。

我的数据看起来像（年龄列）

	Age
0	31.0
1	79.0
2	53.0
3	40.0
4	55.0
	...
44872	NaN
44873	NaN
44874	NaN
44875	NaN
44876	NaN

numpy 和 pandas 有没有可能混在一起？我用它把我的数据分成训练和测试：

from sklearn.model_selection import train_test_split

dep_var = ['is_overdue']
features = model_data2.columns
features = features.drop(dep_var)

print(features)

X = model_data2[features].values
Y = model_data2[dep_var].values

split_test_size = 0.30

X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=split_test_size, random_state=42) 

非常感谢您的帮助。

Answer 1

因为你想用平均值替换 0，你必须用 0 填充 NaN：

fill_0_with_mean = SimpleImputer(missing_values=0, strategy='mean')
X_train['Age'] = fill_0_with_mean.fit_transform(X_train['Age'].fillna(0))