table of Contents
Use sklearn in Python to achieve the above formula calculation
Take the classification problem as an example
| Label 1 | Label 0 | |
| Prediction 1 | TP | FP |
| Prediction 0 | FN | TN |
1. TP (True Positive) true example : indicates that the prediction is true, and the actual is also true
2. FP (False Positive) false positive example : indicates that the prediction is true but the actual is false
3. TN (True Negetive) example : indicates that the prediction is false, but the actual is false
4. FN (False Negetive) false negative example : indicates that the prediction is false but the actual is true
1. Accuracy
That is , the ratio at which all predictions are correct :
2. Precision
The precision rate, the ratio of the correct prediction being positive to the total prediction being positive, and the total prediction being the ratio of the actual label being 1:
3. Recall
Recall rate, the correct prediction is the proportion of all positive samples . The actual label is the ratio of 1 which is correctly predicted to be 1:
4.F值(F-measure)
F-measure is the weighted harmonic average of Precision and Recall:
When the time
Use sklearn in Python to achieve the above formula calculation
method one
# 构建混淆矩阵
from sklearn.metrics import confusion_matrix
confusion_matrix(y_test_labels, y_pred_labels)
# 精准率与召回率
from sklearn.metrics import accuracy_score, precision_score, recall_score
print(accuracy_score(y_test_labels, y_pred_labels))
print('-========')
print(precision_score(y_test_labels, y_pred_labels))
print('-========')
print(recall_score(y_test_labels, y_pred_labels))
# f1 score
from sklearn.metrics import f1_score
f1_score(y_test_labels, y_pred_labels)
Method Two
from sklearn.metrics import precision_recall_fscore_support
precision, recall, f1score, support = precision_recall_fscore_support(y_true, y_pred, beta=1.0, labels=None,
pos_label=1, average=None, warn_for=(‘precision’, ’recall’, ’f-score’), sample_weight=None)