Pipe Line in ML

 

Step 1: Define Preprocessing Steps¶

Similar to how a pipeline bundles together preprocessing and modeling steps, we use the ColumnTransformer class to bundle together different preprocessing steps. The code below:

• imputes missing values in numerical data, and
• imputes missing values and applies a one-hot encoding to categorical data.

from sklearn.compose import ColumnTransformer

from sklearn.pipeline import Pipeline

from sklearn.impute import SimpleImputer

from sklearn.preprocessing import OneHotEncoder

from sklearn.ensemble import RandomForestRegressor

from sklearn.metrics import mean_absolute_error

# Preprocessing for numerical data

numerical_transformer = SimpleImputer(strategy='constant')

# Preprocessing for categorical data

categorical_transformer = Pipeline(steps=[

    ('imputer', SimpleImputer(strategy='most_frequent')),

    ('onehot', OneHotEncoder(handle_unknown='ignore'))

])

# Bundle preprocessing for numerical and categorical data

preprocessor = ColumnTransformer(

    transformers=[

        ('num', numerical_transformer, numerical_cols),

        ('cat', categorical_transformer, categorical_cols)

    ])

# Define model

model = RandomForestRegressor(n_estimators=100, random_state=0)

# Bundle preprocessing and modeling code in a pipeline

clf = Pipeline(steps=[('preprocessor', preprocessor),

                      ('model', model)

                     ])

# Preprocessing of training data, fit model 

clf.fit(X_train, y_train)

# Preprocessing of validation data, get predictions

preds = clf.predict(X_valid)

print('MAE:', mean_absolute_error(y_valid, preds))