How to choose the optimal threshold for imbalanced classification

threshold_roc.py

# Create the ROC curve
fpr, tpr, thresholds = roc_curve(y_test, y_pred)

# Plot the ROC curve
df_fpr_tpr = pd.DataFrame({'FPR':fpr, 'TPR':tpr, 'Threshold':thresholds})
df_fpr_tpr.head()

# Create the data viz
plotnine.options.figure_size = (8, 4.8)
(
    ggplot(data = df_fpr_tpr)+
    geom_point(aes(x = 'FPR',
                   y = 'TPR'),
               size = 0.4)+
    geom_line(aes(x = 'FPR',
                  y = 'TPR'))+
    labs(title = 'ROC Curve')+
    xlab('False Positive Rate')+
    ylab('True Positive Rate')+
    theme_minimal()
)