missing_values_with_knnimputer.py

# import packages
import pandas as pd
import numpy as np
from sklearn.impute import KNNImputer

# create a sample dataframe
my_df = pd.DataFrame({"A": [1,2,3,4,5],
                      "B": [1,1,3,3,4],
                      "C": [1,2,9,np.nan,20]})


# define the package into a variable
knn_imputer = KNNImputer()


# define the package into a variable with a specific parameter 
# the default n_neighbors is 5
knn_imputer = KNNImputer(n_neighbors = 1)
knn_imputer = KNNImputer(n_neighbors = 2)

# define the package into a variable with the weights parameter
# the default value of weights = uniform
knn_imputer = KNNImputer(n_neighbors = 2, weights = "distance")
knn_imputer.fit_transform(my_df)


# change the datatype to dataframe 
my_df1 = pd.DataFrame(knn_imputer.fit_transform(my_df), columns = my_df.columns)