2020-08-24绘制ROC PR曲线核心方法总结，计算AUC核心方法

#1 绘制ROC AUC PR曲线

#核心方法

roc_curve(y_test, preds) # preds为概率形式

#source code
import matplotlib.pyplot as plt
plt.figure(figsize=(12,8))
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.title('ROC Curve for different classifiers')

plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')

labels = []
preds = model.predict(val_data_1,val_data_2)
y_test = labels_val
print("preds=",preds)
print("y_test=",labels_val)
fpr, tpr, _ = roc_curve(y_test, preds)
correct_prediction = np.equal(np.round(preds), y_test)
print("准确率=",np.mean(correct_prediction))
roc_auc = auc(fpr, tpr)
plt.plot(fpr, tpr, lw=lw, color=colors[idx])
labels.append(" AUC = {}".format(np.round(roc_auc, 4)))
plt.legend(['random AUC = 0.5'] + labels)

#2 绘制PR曲线

#核心方法

precision_recall_curve(y_test, preds) #preds为概率形式

#绘制PR曲线
# precision_recall_curve 评价

80858/80858 [==============================] - 71s 879us/step
preds= [[0.47779566]
 [0.10188359]
 [0.9416214 ]
 ...
 [0.35241508]
 [0.08200879]
 [0.8329295 ]]
y_test= [1 0 1 ... 1 0 1]
准确率= 0.5920709617260591
AUC1 is  0.8846388994626855
AUC2 is  0.8846388994626855

Out[104]:

<matplotlib.legend.Legend at 0x7f16bfbd4e80>

In [105]:

#绘制PR曲线

# precision_recall_curve 评价

from sklearn.metrics import precision_recall_curve

pr, re, _ = precision_recall_curve(y_test, preds)

plt.figure(figsize=(12,8))

plt.plot(re, pr)

plt.title('PR Curve (AUC {})'.format(auc(re, pr)))

plt.xlabel('Recall')

plt.ylabel('Precision')

Out[105]:

Text(0, 0.5, 'Precision')

In [108]:

#灵敏度计算

print("sens=",tpr)

sense = tpr

specificity = 1-fpr

print("specificity 1-fpr=",1-fpr)

plt.figure(figsize=(12,8))

plt.plot(specificity, sense)

plt.title('sensitivity-specificity curve')

plt.xlabel('specificity ')

plt.ylabel('sensitivity')

sens= [0.         0.00198908 0.00212393 ... 1.         1.         1.        ]
specificity 1-fpr= [1.00000000e+00 1.00000000e+00 1.00000000e+00 ... 1.17196656e-04
 3.90655520e-05 0.00000000e+00]

Out[108]: