In sklearn, precision_score, recall_recore, f1_score of two categories

precision_score: precision rate, precision rate
$$P=\frac{TP}{TP+FP}$$

#	假设二分类标签为1，2
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
precision_score(y_true, y_pred, average="binary", pos_label=1)
# pos_label设置为1，代表标签为1的样本是正例，标签为2的样本是负例。

accuracy_score: accuracy rate
$$Acc=\frac{TP+TN}{TP+FP+TN+FN}$$

#	假设二分类标签为1，2
accuracy_score(y_true, y_pred)

recall_score: recall rate, recall rate
$$R=\frac{TP}{TP+FN}$$

#	假设二分类标签为1，2
recall_score(y_true, y_pred, average="binary", pos_label=1)

f1_score: F1值
$$F1=\frac{2\ast P\ast R}{P+R}$$

#	假设二分类标签为1，2
f1_score(y_true, y_pred, average="binary", pos_label=1)

Attachment: Confusion matrix of two categories

reality	forecast result
	Positive example	Counterexample
Positive example	TP (real case)	FN (false counterexample)
Counterexample	FP (false positive)	TN (True Negative Example)