def cal_tag(result_path):
'''
Calculate evaluation for each tag separately
:param result_path: output file path
:return: dictionary of evaluation results for each tag
'''
evaluation_result = {}
data = get_full(result_path)
for tag in tags:
TP = cal_tp(data, tag)
FN = cal_fn(data, tag)
FP = cal_fp(data, tag)
TN = cal_tn(data, tag)
precision = TP / (TP + FP)
recall = TP / (TP + FN)
accuracy = (TP + TN) / (TP + TN + FN + FP)
f1 = 2 * precision * recall / (precision + recall)
print('tag: ' + tag)
print('precision: ' + str(precision) + '; recall: ' + str(recall) +
'; accurary: ' + str(accuracy) + '; f1: ' + str(f1))
evaluation_result[tag] = {
'precision': precision,
'recall': recall,
'accuracy': accuracy,
'f1': f1
}
return evaluation_result
def cal_label(result_path):
'''
Calculate evaluation on whole label set. Results for each label.
:param result_path: output file path
:return:
'''
data = get_full(result_path)
evaluation_result = {}
for label in labels:
tags = ['B-' + label, 'I-' + label]
TP = cal_tp(data, tags[0]) + cal_tp(data, tags[1])
FN = cal_fn(data, tags[0]) + cal_fn(data, tags[1])
FP = cal_fp(data, tags[0]) + cal_fp(data, tags[1])
TN = cal_tn(data, tags[0]) + cal_tn(data, tags[1])
print(tags)
print(TP, FN, FP, TN)
if TP + FP != 0:
precision = TP / (TP + FP)
else:
precision = 0
if TP + FN != 0:
recall = TP / (TP + FN)
else:
recall = 0
if (TP + TN + FN + FP) != 0:
accuracy = (TP + TN) / (TP + TN + FN + FP)
else:
accuracy = 0
if (precision + recall) != 0:
f1 = 2 * precision * recall / (precision + recall)
else:
f1 = 0
print(label)
print('precision: ' + str(precision) + '; recall: ' + str(recall) +
'; accurary: ' + str(accuracy) + '; f1: ' + str(f1))
evaluation_result[label] = {
'precision': precision,
'recall': recall,
'accuracy': accuracy,
'f1': f1
}
return evaluation_result
def cal_total(result_path):
'''
Calculate evaluation on whole set
:param result_path: output file path
:return:
'''
data = get_full(result_path)
TP = 0
FN = 0
FP = 0
TN = 0
for tag in tags:
_TP = cal_tp(data, tag)
_FN = cal_fn(data, tag)
_FP = cal_fp(data, tag)
_TN = cal_tn(data, tag)
TP += _TP
FN += _FN
FP += _FP
TN += _TN
precision = TP / (TP + FP)
recall = TP / (TP + FN)
accuracy = (TP + TN) / (TP + TN + FN + FP)
f1 = 2 * precision * recall / (precision + recall)
print('precision: ' + str(precision) + '; recall: ' + str(recall) +
'; accurary: ' + str(accuracy) + '; f1: ' + str(f1))
evaluation_result = {
'precision': precision,
'recall': recall,
'accuracy': accuracy,
'f1': f1
}
return evaluation_result
def get_prediction_data(prediction_result_path):
data = get_full(prediction_result_path)
return data[2]