Function get_survival_data_for_BS

utils.py:185–196

(df, time_col_name, censorship_col_name='censorship')

    print(f"Model has {n_trainable_params} parameters")

def get_survival_data_for_BS(df, time_col_name, censorship_col_name='censorship'):
    # To compute one survival metric the Brier score (BS), you need a specific format of censorship and times 
    # This is to estimate the censoring distribution from. 
    # A structured array containing the binary event indicator as first field (1 event occured; 0 censored), and time of event or time of censoring as second field.
    
    val = df[[censorship_col_name, time_col_name]].values
    max_time = df[time_col_name].max()

    y = np.empty(len(df), dtype=[('cens', '?'), ('time', '<f8')])
    for i in range(len(df)):
        y[i] = tuple((bool(1-val[i][0]),val[i][1])) # Note that we take the uncensorship status.
    return max_time, y

def get_bins_time_value(df, n_bins, time_col_name, label_time_col_name='disc_label', censorship_col_name='censorship'):
    # Retrieve the values of each bin from the dataset.