Draw multiple ROC curves on a graph

In order to show the experimental effect ,ROC Curves can also be more intuitive and beautiful . So I want to draw ROC curve . Here are two ways ：1） Just draw one ROC curve ,2） Multiple ROC The curve is shown on a graph .

notes ： I already have y-pred, So just use it directly , There is no need to calculate y_pred.

1 Draw a picture ROC curve

from sklearn.metrics import roc_curve, auc
import matplotlib.pyplot as plt

#  Set up here lable It's really in there 1.
fpr, tpr, thersholds = roc_curve(y_label, y_pred, pos_label=1)

for i, value in enumerate(thersholds):
    print("%f %f %f" % (fpr[i], tpr[i], value))

roc_auc = auc(fpr, tpr)
plt.figure(figsize=(10, 10), dpi=100)
plt.plot(fpr, tpr, 'k--', label='ROC (area = {0:.2f})'.format(roc_auc), lw=2)

plt.xlim([-0.05, 1.05])  #  Set up x、y The upper and lower limits of the axis , So as not to coincide with the edge , Better observe the whole of the image 
plt.ylim([-0.05, 1.05])
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')  #  Can use Chinese , But you need to import some libraries that are Fonts 
plt.title('ROC Curve')
plt.legend(loc="lower right")
plt.show()

 2 Will be multiple ROC The curve is shown on a graph

Define drawing function

def multi_models_roc(names, sampling_methods, colors, y_label, save=True, dpin=100):
    """
     Combine multiple machine models roc Graph output to a graph 
    
    Args:
        names: list,  Names of multiple models 
        sampling_methods: list,  Instantiated objects of multiple models 
        save:  Choose whether to save the results （ The default is png Format ）
        
    Returns:
         Return the image object plt
    """
    plt.figure(figsize=(10, 10), dpi=dpin)

    for (name, y_pred, colorname) in zip(names, sampling_methods, colors):
        
      
        fpr, tpr, thresholds = roc_curve(y_label, y_pred, pos_label=1)
        
        plt.plot(fpr, tpr, lw=5, label='{} (AUC={:.3f})'.format(name, auc(fpr, tpr)),color = colorname)
        plt.plot([0, 1], [0, 1], '--', lw=5, color = 'grey')
        plt.axis('square')
        plt.xlim([0, 1])
        plt.ylim([0, 1])
        plt.xlabel('False Positive Rate',fontsize=20)
        plt.ylabel('True Positive Rate',fontsize=20)
        plt.title('ROC Curve',fontsize=25)
        plt.legend(loc='lower right',fontsize=20)

    if save:
        plt.savefig('multi_models_roc.png')
        
    return plt

call multi_models_roc

names = ['UTrans_Mem',
         'Unet_Mem',
         ]

# sampling_methods Li Wei  y_pred.
sampling_methods = [preds_UT_mem,
                    preds_U_mem,
                   ]
#color:'crimson','orange','gold','mediumseagreen','steelblue', 'mediumpurple' 
colors = ['crimson',
          'orange',
         ]

#ROC curves
train_roc_graph = multi_models_roc(names, sampling_methods, colors, label[0], save = True)
train_roc_graph.savefig('ROC_Train_all.png')

 

The following links are more detailed .

(56 Bar message ) python Combine the... Of multiple models ROC The curve is drawn in a graph （ With legend ）_ Liunainainai's blog -CSDN Blog _pytorch Many classification roc Curve drawing