In [20]:

    

from sklearn import datasets as d
import numpy as np
from sklearn import preprocessing


    

In [2]:

    

iris = d.load_iris()
iris_x = iris.data


    

In [3]:

    

masking_array = np.random.binomial(1,.25, iris_x.shape).astype(bool)


    

In [17]:

    

iris_x[masking_array] = np.nan


    

In [18]:

    

masking_array[:5]


    

    
Out[18]:





array([[ True,  True,  True, False],
       [ True,  True,  True, False],
       [False,  True, False, False],
       [False, False, False,  True],
       [False,  True, False,  True]], dtype=bool)

In [19]:

    

iris_x[:5]


    

    
Out[19]:





array([[ nan,  nan,  nan,  0.2],
       [ nan,  nan,  nan,  0.2],
       [ 4.7,  nan,  1.3,  0.2],
       [ 4.6,  3.1,  1.5,  nan],
       [ 5. ,  nan,  1.4,  nan]])

In [21]:

    

# could pass strategy='median' or 'most_frequent'
impute = preprocessing.Imputer()


    

In [23]:

    

iris_x_prime = impute.fit_transform(iris_x)


    

In [24]:

    

iris_x_prime[:5]


    

    
Out[24]:





array([[ 5.8754386 ,  3.07232143,  3.75203252,  0.2       ],
       [ 5.8754386 ,  3.07232143,  3.75203252,  0.2       ],
       [ 4.7       ,  3.07232143,  1.3       ,  0.2       ],
       [ 4.6       ,  3.1       ,  1.5       ,  1.14504505],
       [ 5.        ,  3.07232143,  1.4       ,  1.14504505]])

In [ ]: