article-identify-hardbounce-ml_2_aux_functions.py

def oversampleSMOTE(X, y):
    '''
    Resample a dataset using SMOTE oversample

    Input:
        X = dataframe with x variables (explanatory variables)
        y = dataframe with y variable (variable to predict)

    Output:
        df[0] = X dataframe resampled
        df[1] = y dataframe resampled
    '''

    from imblearn.over_sampling import SMOTE
    import pandas as pd

    sm = SMOTE(random_state=123)
    X_resampled, y_resampled = sm.fit_resample(X, y.ravel())

    # Get column names
    X_cols = X.columns.values
    y_cols = [y.name]

    return pd.DataFrame(X_resampled, columns=X_cols) , pd.DataFrame(y_resampled, columns=y_cols)

article-identify-hardbounce-ml_2_remove-imbalance.ipynb