Fatal编程技术网
  • python/
  • Python 我无法理解代码中计算的FPR、TPR、阈值、ROC值的含义（直觉）？

Python 我无法理解代码中计算的FPR、TPR、阈值、ROC值的含义（直觉）？

python numpy machine-learning scikit-learn

Python 我无法理解代码中计算的FPR、TPR、阈值、ROC值的含义（直觉）？,python,numpy,machine-learning,scikit-learn,roc,Python,Numpy,Machine Learning,Scikit Learn,Roc,我想为我的分类模型绘制ROC曲线。因为我是新手，所以我读了一些关于它的文章，看到了一些帖子，我创建了一条roc曲线 我的数据是这种类型的： print(Y.shape) print(predictions.shape) print(Y) print(predictions) (1, 400) (1, 400) [[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

我想为我的分类模型绘制ROC曲线。因为我是新手，所以我读了一些关于它的文章，看到了一些帖子，我创建了一条roc曲线

我的数据是这种类型的：

print(Y.shape)
print(predictions.shape)
print(Y)
print(predictions)

(1, 400)
(1, 400)
[[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
  1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
  1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
  1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
  1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
  1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
  1 1 1 1]]
[[0 1 1 1 0 1 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
  0 0 0 0 0 0 0 0 0 0 1 0 1 1 1 0 1 1 1 1 0 1 1 1 0 1 0 0 1 1 1 1 1 1 0 0
  1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
  1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 0 1
  0 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
  1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 1 1 1
  1 1 1 1 1 0 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 1 0
  0 0 1 0]]
实施该准则后：

from sklearn.metrics import precision_score
print('Precsion score: '+ str(precision_score(Y.ravel(), predictions.ravel())))

from sklearn.metrics import recall_score
print('Recall score: '+ str(recall_score(Y.ravel(), predictions.ravel())))

from sklearn.metrics import f1_score
print('F1 score: '+ str(f1_score(Y.ravel(), predictions.ravel())))

from sklearn.metrics import roc_auc_score, auc, roc_curve
print('ROC score: ' + str(roc_auc_score(Y.ravel(), predictions.ravel())))

from sklearn.metrics import confusion_matrix
print('Confusion matrix: ')
print(confusion_matrix(Y.ravel(), predictions.ravel()))

fpr = dict()
tpr = dict()
threshold = dict()
roc_auc = dict()

for i in range(2):
    fpr[i], tpr[i], threshold[i] = roc_curve(Y.ravel(), predictions.ravel())
    roc_auc[i] = auc(fpr[i], tpr[i])
print(fpr, tpr, threshold, roc_auc)

plt.figure()
plt.plot(fpr[1], tpr[1])
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.title('Receiver operating characteristic(ROC Curve)')
输出：

Precsion score: 0.9179487179487179
Recall score: 0.895
F1 score: 0.9063291139240507
ROC score: 0.9075
Confusion matrix: 
[[184  16]
 [ 21 179]]
{0: array([0.  , 0.08, 1.  ]), 1: array([0.  , 0.08, 1.  ])} {0: array([0.   , 0.895, 1.   ]), 1: array([0.   , 0.895, 1.   ])} {0: array([2, 1, 0]), 1: array([2, 1, 0])} {0: 0.9075, 1: 0.9075}
Text(0.5, 1.0, 'Receiver operating characteristic(ROC Curve)')

我不明白为什么要使用循环？我可以看到每行计算的FPR、TPR、Threshold和roc_auc的三个值。我确实读到过roc_曲线以概率作为目标分数（我会继续研究）。但是，我无法从输入（1400）维度数据了解这些数组是如何计算的

提前感谢。

我也不明白您为什么使用循环，因为您可以通过删除并调整代码来获得与代码相同的功能：

import matplotlib.pyplot as plt
from sklearn.metrics import precision_score
from sklearn.metrics import recall_score
from sklearn.metrics import f1_score
from sklearn.metrics import roc_auc_score, auc, roc_curve
from sklearn.metrics import confusion_matrix

Y=[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1]

predictions=[0, 1, 1, 1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0, 1, 1,
       1, 1, 0, 1, 1, 1, 0, 1, 0, 0, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 0,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0,
       0, 0, 1, 0]

print('Precsion score: '+ str(precision_score(Y, predictions)))
print('Recall score: '+ str(recall_score(Y, predictions)))
print('F1 score: '+ str(f1_score(Y, predictions)))
print('ROC score: ' + str(roc_auc_score(Y, predictions)))
print('Confusion matrix: ')
print(confusion_matrix(Y, predictions))

fpr, tpr, threshold = roc_curve(Y, predictions)
roc_auc = auc(fpr, tpr)

print(fpr, tpr, threshold, roc_auc)

plt.figure()
plt.plot(fpr, tpr)
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.title('Receiver operating characteristic(ROC Curve)')
它产生以下输出：

Precsion score: 0.9179487179487179
Recall score: 0.895
F1 score: 0.9063291139240507
ROC score: 0.9075
Confusion matrix: 
[[184  16]
 [ 21 179]]
[0.   0.08 1.  ] [0.    0.895 1.   ] [2 1 0] 0.9075

您的400个数据点用于计算ROC曲线，但可视化中只有三个数据点，因为您的数据中只有两个唯一的值（0和1）

引用以下的答案：

点的数量取决于中唯一值的数量 输入。由于输入向量只有2个唯一值，因此函数 提供正确的输出

正如链接的SO答案明确指出的那样，绘制ROC曲线需要概率预测，而不是像这里这样的硬类0/1。提醒一下，ROC曲线需要概率预测，而不是硬类0/1应该在答案中。