forked from dapowan/LIMU-BERT-Public
-
Notifications
You must be signed in to change notification settings - Fork 0
/
statistic.py
59 lines (50 loc) · 2.22 KB
/
statistic.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
# !/usr/bin/env python
# -*- coding: utf-8 -*-
# @Time : 14/1/2021
# @Author : Huatao
# @Email : [email protected]
# @File : statistic.py
# @Description :
import numpy as np
from sklearn import metrics
from sklearn.metrics import f1_score
from plot import plot_matrix
def stat_acc_f1(label, results_estimated):
# label = np.concatenate(label, 0)
# results_estimated = np.concatenate(results_estimated, 0)
label_estimated = np.argmax(results_estimated, 1)
f1 = f1_score(label, label_estimated, average='macro')
acc = np.sum(label == label_estimated) / label.size
return acc, f1
def stat_acc_f1_dual(label, results_estimated):
label = np.concatenate(label, 0)
results_estimated = np.concatenate([t[1] for t in results_estimated], 0)
label_estimated = np.argmax(results_estimated, 1)
f1 = f1_score(label, label_estimated, average='macro')
acc = np.sum(label == label_estimated) / label.size
return acc, f1
def stat_results(label, results_estimated):
label_estimated = np.argmax(results_estimated, 1)
f1 = f1_score(label, label_estimated, average='macro')
acc = np.sum(label == label_estimated) / label.size
matrix = metrics.confusion_matrix(label, label_estimated) #, normalize='true'
return acc, matrix, f1
def stat_acc_f1_tpn(label, label_estimated, task_num=5, threshold=0.5):
label_new = []
label_estimated_new = []
for i in range(label.size):
if label[i] == 0:
label_new.append(np.zeros((task_num, 1)))
label_estimated_new_temp = np.zeros((task_num, 1))
label_estimated_new_temp[label_estimated[i, :] > threshold] = 1
label_estimated_new.append(label_estimated_new_temp)
else:
label_new.append(np.ones((1, 1)))
label_estimated_new_temp = np.zeros((1, 1))
label_estimated_new_temp[label_estimated[i, label[i] - 1] > threshold] = 1
label_estimated_new.append(label_estimated_new_temp)
label_new = np.concatenate(label_new, 0)[:, 0]
label_estimated_new = np.concatenate(label_estimated_new, 0)[:, 0]
f1 = f1_score(label_new, label_estimated_new, average='macro')
acc = np.sum(label_new == label_estimated_new) / label_new.size
return acc, f1