MAR univariate example

A basic example of generate artificial missing data under Missing at Random (MAR) mechanism with mdatagen library with the Iris dataset from scikit-learn. The observed feature is "petal length" and the feature that will receive the missing values is "petal width". The simulated missing rate is 12%. The method to choose missing values is random.

In [1]:

# Import the libraries
import pandas as pd
from sklearn.datasets import load_iris
from mdatagen.univariate.uMAR import uMAR

# Load the data
iris = load_iris()
iris_df = pd.DataFrame(data=iris.data, columns=iris.feature_names)

X = iris_df.copy()   # Features
y = iris.target    # Label values

# Create a instance with missing rate equal to 12% in dataset under MAR mechanism
generator = uMAR(X=X, 
                 y=y, 
                 missing_rate=12, 
                 x_miss='petal width (cm)',
                 x_obs='petal length (cm)')

# Generate the missing data under MNAR mechanism
generate_data = generator.random()
print(generate_data.isna().sum())

# Import the libraries import pandas as pd from sklearn.datasets import load_iris from mdatagen.univariate.uMAR import uMAR # Load the data iris = load_iris() iris_df = pd.DataFrame(data=iris.data, columns=iris.feature_names) X = iris_df.copy() # Features y = iris.target # Label values # Create a instance with missing rate equal to 12% in dataset under MAR mechanism generator = uMAR(X=X, y=y, missing_rate=12, x_miss='petal width (cm)', x_obs='petal length (cm)') # Generate the missing data under MNAR mechanism generate_data = generator.random() print(generate_data.isna().sum())

sepal length (cm)     0
sepal width (cm)      0
petal length (cm)    38
petal width (cm)      0
target                0
dtype: int64

In [ ]: