基于ELECTRA-CRF的电信网络诈骗案件文本命名实体识别模型

doi:10.3969/j.issn.1671-1122.2021.06.008

信息网络安全 ›› 2021, Vol. 21 ›› Issue (6): 63-69.doi: 10.3969/j.issn.1671-1122.2021.06.008

基于ELECTRA-CRF的电信网络诈骗案件文本命名实体识别模型

丁家伟¹, 刘晓栋²()

1. 中国人民公安大学侦查学院,北京 100038
2. 中国人民公安大学治安与交通管理学院,北京 100038

收稿日期:2021-04-29 出版日期:2021-06-10 发布日期:2021-07-01
通讯作者: 刘晓栋 E-mail:liuxiaodong@ppsuc.edu.cn
作者简介:丁家伟（1997—）,男,山东,硕士研究生,主要研究方向为刑事侦查学|刘晓栋（1988—）,男,山东,讲师,博士,主要研究方向为公安大数据、应急管理
基金资助:
国家重点研发计划(2020YFC1522600)

Named Entity Recognition Model of Telecommunication Network Fraud Crime Based on ELECTRA-CRF

DING Jiawei¹, LIU Xiaodong²()

1. College of Investigation, People’s Public Security University of China, Beijing 100038, China;
2. College of Public Security and Traffic Management, People’s Public Security University of China, Beijing 100038, China;

Received:2021-04-29 Online:2021-06-10 Published:2021-07-01
Contact: LIU Xiaodong E-mail:liuxiaodong@ppsuc.edu.cn

摘要/Abstract

摘要：

文章提出一种基于ELECTRA-CRF的电信网络诈骗案件文本命名实体识别模型。该模型首先将标注后的语料输入ELECTRA模型,得到以字为颗粒度的状态转移特征;然后由CRF模型计算转移分数,判断当前位置与其相邻位置字符的实体标注组合;最后将该模型与BERT-CRF模型、RoBERTa-CRF模型进行对比。实验结果表明,文中模型在运算效率上明显优于其他两种深度学习模型,且准确度、召回率和调和平均值并未有太大损失,可以很好地应用于电信网络诈骗案件的命名实体识别中。

关键词: 命名实体识别, ELECTRA模型, 电信网络诈骗

Abstract:

This paper proposes a text named entity recognition model of telecommunication network fraud crimes based on ELECTRA-CRF. Firstly, the annotated corpus is input into ELECTRA model to obtain the state transition features with Chinese characters as granularity. And then CRF model is used to calculate the transfer score to determine the entity label group of the character at the current position and its adjacent position. Finally, the BERT-CRF model and RoBERTa-CRF model are compared through experiments. The experimental results show that the text named entity recognition model proposed in this paper based on ELECTRA-CRF is significantly better than the other two deep learning models in operation efficiency, and the loss of the accuracy, recall rate and reconciliation average are very small. It can be well applied to the named entity recognition of telecommunication network fraud crimes.

Key words: named entity recognition, ELECTRA model, telecommunication network fraud crime

中图分类号:

TP309

丁家伟, 刘晓栋. 基于ELECTRA-CRF的电信网络诈骗案件文本命名实体识别模型[J]. 信息网络安全, 2021, 21(6): 63-69.

DING Jiawei, LIU Xiaodong. Named Entity Recognition Model of Telecommunication Network Fraud Crime Based on ELECTRA-CRF[J]. Netinfo Security, 2021, 21(6): 63-69.

图/表 7

图1

图2

图3

表1

表2

图4

表3

参考文献 16

[1]	CHEN Jian, HE Tao, WEN Yingyou, et al. Entity Recognition Method for Judicial Documents Based on BERT Model[J]. Journal of Northeastern University: Natural Science, 2020,41(10):1382-1387.
	陈剑, 何涛, 闻英友, 等. 基于BERT模型的司法文书实体识别方法[J]. 东北大学学报:自然科学版, 2020,41(10):1382-1387.
[2]	ZHOU Xiaolei, ZHAO Xuejiao, LIU Tangliang, et al. Named Entity Recognition Method of Judgment Documents with SVM-BiLSTM-CRF[J]. Computer Systems and Applications, 2019,28(1):245-250.
	周晓磊, 赵薛蛟, 刘堂亮, 等. 基于SVM-BiLSTM-CRF模型的财产纠纷命名实体识别方法[J]. 计算机系统应用, 2019,28(1):245-250.
[3]	LI Shuqin, LIU Zhao, SHI Yuntao. Research on Named Entity Recognition Method for Food Safety Judgment Documents[J]. Computer Era, 2020(10):1-3, 8.
	李书钦, 刘召, 史运涛. 面向食品安全裁判文书的命名实体识别方法研究[J]. 计算机时代, 2020(10):1-3,8.
[4]	HUANG Han, WANG Hongyu, WANG Xiaoguang. Conditional Random Field Model Combined with Active Learning for Automatic Recognition of Legal Terms[J]. Data Analysis and Knowledge Discovery, 2019,3(6):66-74.
	黄菡, 王宏宇, 王晓光. 结合主动学习的条件随机场模型用于法律术语的自动识别[J]. 数据分析与知识发现, 2019,3(6):66-74.
[5]	WANG Liu. Design and Implementation of Telecommunication Fraud Semantic Pattern Extraction System[D]. Beijing: Beijing University of Posts and Telecommunications, 2019.
	王刘. 电信诈骗语义模式抽取系统的设计与实现[D]. 北京:北京邮电大学, 2019.
[6]	PU Liming. Design and Implementation of Semantic Classification System for Telecommunication Fraud[D]. Beijing: Beijing University of Posts and Telecommunications, 2019.
	蒲黎明. 电信诈骗语义分类系统的设计与实现[D]. 北京:北京邮电大学, 2019.
[7]	WANG Yue, WANG Mengxuan, ZHANG Sheng, et al. Alarm Text Named Entity Recognition Based on BERT[J]. Journal of Computer Applications, 2020,40(2):535-540.
	王月, 王孟轩, 张胜, 等. 基于BERT的警情文本命名实体识别[J]. 计算机应用, 2020,40(2):535-540.
[8]	ZHOU Zhining, WANG Binjun, ZHAI Yiming, et al. Spam Filtering Model Based on ALBERT Dynamic Word Vector[J]. Netinfo Security, 2020,20(9):107-111.
	周枝凝, 王斌君, 翟一鸣, 等. 基于ALBERT动态词向量的垃圾邮件过滤模型[J]. 信息网络安全, 2020,20(9):107-111.
[9]	CLARK K, LUONG M T, LE Q V, et al. Pre-training Text Encoders as Discriminators Rather Than Generators[EB/OL]. https://arxiv.org/abs/2003.10555, 2020-12-11.
[10]	WALLACH H M. Conditional Random Fields: An Introduction[EB/OL]. https://www.researchgate.net/publication/32205774_Conditional_Random_Fields_An_Introduction, 2020 -12-12.
[11]	LIU Weijun. On Case Structure as a Fundamental Theory of Criminal Investigation[J]. Journal of Public Security Science, 2020,3(3):104-121, 124.
	刘为军. 论作为侦查学基础理论的案件结构理论[J]. 公安学研究, 2020,3(3):104-121,124.
[12]	GUL K S Q, YIN Jize, PAN Limin, et al. Research on the Algorithm of Named Entity Recognition Based on Deep Neural Network[J]. Netinfo Security, 2017,17(10):29-35.
	GUL K S Q, 尹继泽, 潘丽敏, 等. 基于深度神经网络的命名实体识别方法研究[J]. 信息网络安全, 2017,17(10):29-35.
[13]	CUI Yiming, CHE Wanxiang, LIU Ting, et al. Revisiting Pre-trained Models for Chinese Natural Language Processing[EB/OL]. https://arxiv.org/abs/2004.13922, 2020-12-13.
[14]	DEVLIN J, CHANG Mingwei, LEE K, et al. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding[EB/OL].https://arxiv.org/abs/1810.04805 , 2020-12-16.
[15]	CUI Yiming, CHE Wanxiang, LIU Ting, et al. Pre-training with Whole Word Masking for Chinese BERT[EB/OL]. , 2020 -12-16.
[16]	LIU Yinhan, OTT M, GOYAL N, et al. Roberta: A robustly Optimized BERT Pretraining Approach[EB/OL]. https://arxiv.org/abs/1907.11692v1, 2020 -12-16.

命名实体类	数量/个	高频实体
ContactWay	1020	手机、电话、微信
ContactNum	265	无
Identity	1392	客服、工作人员
Trick	1382	六合彩、安全账户
Payment	1020	银行转账、支付宝
PayName	592	无
Amount	1352	无

类别	具体信息
操作系统	Windows 10
CPU	AMD Ryzen 7 4800H
RAM	16 GB
TensorFlow	1.14
Python	3.7.8

模型实体类	ELECTRA-CRF			BERT-CRF			RoBERTa-CRF
模型实体类	P	R	F	P	R	F	P	R	F
ContactWay	0.85	0.89	0.87	0.90	0.94	0.92	0.86	0.97	0.91
ContactNum	0.76	0.95	0.84	0.86	0.95	0.90	0.90	0.95	0.93
Identity	0.85	0.91	0.88	0.81	0.90	0.85	0.76	0.87	0.81
Trick	0.66	0.85	0.75	0.71	0.88	0.79	0.66	0.87	0.75
Payment	0.70	0.85	0.77	0.67	0.87	0.75	0.57	0.74	0.64
Amount	0.95	0.99	0.97	0.94	0.99	0.96	0.95	0.99	0.97
PayName	0.97	0.97	0.97	0.98	0.98	0.98	0.99	0.97	0.98
平均值	0.80	0.90	0.85	0.80	0.92	0.86	0.77	0.90	0.83

基于ELECTRA-CRF的电信网络诈骗案件文本命名实体识别模型

Named Entity Recognition Model of Telecommunication Network Fraud Crime Based on ELECTRA-CRF

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 7

参考文献 16

相关文章 2

编辑推荐

Metrics

本文评价

[1]	GULKhanSafiQamas, 尹继泽, 潘丽敏, 罗森林. 基于深度神经网络的命名实体识别方法研究[J]. 信息网络安全, 2017, 17(10): 29-35.
[2]	郝文江, 徐丽萍, 姜劲蕾, 李翠翠. 电信网络诈骗犯罪治理技术研究[J]. 信息网络安全, 2016, 16(9): 213-217.