The SciKitlearn’s Imputer (sklearn.preprocessing.Imputer) class, widely used for Imputing, cleaning up and manipulating data sets, especially where there are missing, data has been deprecated from SciKitlearn version 0.22.

sklearn.impute.SimpleImputer is now the preferred class, it is similar to sklearn.preprocessing.Imputer but more succinct.

Here is how to use it

Using the following data set (below) as example (stored as data.csv)

Country	Age	Salary	Purchase
USA	25	60000	Yes
USA	30	75000	No
Canada	42	40000	Yes
USA		30000	Yes
Canada	42		No
…	…	…	…

First we will import numpy (for numpy array functions) and pandas

import numpy as np
import pandas as pd

We will then import the entire data frame (df) from the data set (data.csv) using pandas.read_csv

import numpy as np
import pandas as pd

df = pd.read_csv('data.csv')

For convenience, Lets identify the dependent column name (title_y)

import numpy as np
import pandas as pd

df = pd.read_csv('data.csv')

title_y = 'Purchased'

Now let create the matrix of features and dependent vector from df

import numpy as np
import pandas as pd

df = pd.read_csv('data.csv')

title_y = 'Purchased'

X = df.drop(columns=[title_y]).values 
y = df['title_y'].values 

now, let’s use the new sklearn.impute.SimpleImputer to fix missing data. In this example, we will be replacing missing values with the mean values.

Here are the steps:

• import SimpleImputer
• create a SimpleImputer instance (imputer), with arguments missing_values=np.nan (to identify missing fields) and strategy=’mean’
• fit missing columns (in our example X[:,1:3]) to the imputer
• transform the columns with missing

import numpy as np
import pandas as pd

df = pd.read_csv('data.csv')

title_y = 'Purchased'

X = df.drop(columns=[title_y]).values 
y = df['title_y'].values 

from sklearn.impute import SimpleImputer 

imputer = SimpleImputer(missing_values=np.nan, strategy='mean')
imputer = imputer.fit(X[:,1:3])
replaced = imputer.transform(X[:,1:3])
X[:,1:3] = replaced

And that’s all!