基于大语言模型与事件演化图谱的诈骗电话识别方法

胡程忆

doi:10.11959/j.issn.1000-0801.DXKX250564

您当前的位置：

首页 >

文章列表页 >

基于大语言模型与事件演化图谱的诈骗电话识别方法

研究与开发 | 更新时间：2026-07-21

- 基于大语言模型与事件演化图谱的诈骗电话识别方法
- Fraud call identification based on large language models and event evolution graphs
- 电信科学 2026年42卷第6期页码：203-216
- 作者机构：
  
  卓望信息技术（北京）有限公司，北京 100072
- 作者简介：
  
  胡程忆（1980- ），男，现就职于卓望信息技术（北京）有限公司，主要从事自然语言处理与安全风控工作。
- 基金信息：
- DOI：10.11959/j.issn.1000-0801.DXKX250564
  中图分类号： TP391
- 收稿：2025-09-22，
  
  修回：2025-12-29，
  
  录用：2026-02-10，
  
  网络首发：2026-07-21，
  
  纸质出版：2026-06-20
- 稿件说明：
移动端阅览
胡程忆.基于大语言模型与事件演化图谱的诈骗电话识别方法[J].电信科学,2026,42(06):203-216.

Hu Chengyi.Fraud call identification based on large language models and event evolution graphs[J].Telecommunications Science,2026,42(06):203-216.
胡程忆.基于大语言模型与事件演化图谱的诈骗电话识别方法[J].电信科学,2026,42(06):203-216. DOI： 10.11959/j.issn.1000-0801.DXKX250564.

Hu Chengyi.Fraud call identification based on large language models and event evolution graphs[J].Telecommunications Science,2026,42(06):203-216. DOI： 10.11959/j.issn.1000-0801.DXKX250564.

摘要

针对电信诈骗话术快速演化、传统方法依赖大量标注样本以及基于大模型的端到端识别存在幻觉风险等问题，提出一种基于大语言模型与事件演化图谱的诈骗电话识别方法。该方法引入单类学习思想，在无需负样本的条件下，利用大语言模型的零样本能力，将主叫话术抽象为标准化的动宾短语事件演化链，通过语义相似度融合构建可增量扩展的事件演化图谱，并采用关键事件节点匹配与连通性检测进行风险判别。实验结果表明，该方法的

1值在2 000条样本规模下达84.84%，在8 000条样本规模下可提升至91.46%，性能随样本规模增长呈现一定的缩放律特征。该方法为识别持续演化的复杂诈骗行为提供了精准且具有扩展性的技术路径。

Abstract

In response to the rapid evolution of fraud call scripts

the reliance on a large number of labeled samples in traditional methods

and the potential hallucination risks in end-to-end recognition based on large models

a fraud call identification method based on large language models and event evolution graphs was proposed. By introducing the concept of one-class learning

the zero-shot capability of large language models was leveraged to abstract caller speech into standardized verb-object phrase event evolution chains without the need for negative samples. An incrementally expandable event evolution graph was constructed through semantic similarity fusion

and risk discrimination was performed using key event node matching and connectivity detection. Experimental results show that the proposed method achieves an

1 score of 84.84% on a sample scale of 2 000 instances

which is further improved to 91.46% on a scale of 8 000 instances

with performance exhibiting certain scaling law characteristics as the sample size increases. The proposed method provides a precise and scalable technical approach for identifying complex and continuously evolving fraudulent behaviors.

关键词

Keywords

references

陈禹橦 . 刑事一体化视野中的跨境电信网络诈骗犯罪疑难问题研究 [J ] . 法律适用 , 2024 ( 10 ): 117 - 132 .

Chen Y T . Study on the difficulties of regulation on cross-border telecom and online fraud crimes in the perspective of criminal integration [J ] . Journal of Law Application , 2024 ( 10 ): 117 - 132 .

单勇 . 《反电信网络诈骗法》的均衡实现: 预防性秩序和正当性权益之间 [J ] . 法制与社会发展 , 2025 , 31 ( 1 ): 40 - 64 .

Shan Y . Balanced realization of anti-telecom and online fraud law: between preventive order and legitimate rights [J ] . Law and Social Development , 2025 , 31 ( 1 ): 40 - 64 .

张照星 , 胡珵璇 , 范英 , 等 . 电信诈骗案件中基于运营商国际漫游话单的嫌疑号码拓展 [J ] . 信息网络安全 , 2021 , 21 ( S1 ): 12 - 16 .

Zhang Z X , Hu C X , Fan Y , et al . The expansion of suspect number based on operator’s international roaming bill in telecom fraud cases [J ] . Netinfo Security , 2021 , 21 ( S1 ): 12 - 16 .

李洋 , 李振华 , 辛显龙 . 基于攻击经济学的移动虚拟运营商诈骗检测 [J ] . 计算机科学 , 2023 , 50 ( 8 ): 260 - 270 .

Li Y , Li Z H , Xin X L . Attack economics based fraud detection for MVNO [J ] . Computer Science , 2023 , 50 ( 8 ): 260 - 270 .

刘晨迪 , 李耀 , 陈佳逸 , 等 . 融合图结构和多通道注意力机制的诈骗电话识别 [J ] . 计算机应用 , 2025 , 45 ( S2 ): 58 - 63 .

Liu C D , Li Y , Chen J Y , et al . Scam call identification by integrating graph structure and multi-channel attention mechanism [J ] . Journal of Computer Applications , 2025 , 45 ( S2 ): 58 - 63 .

周俊杰 , 许鸿奎 , 卢江坤 , 等 . 引入位置信息和Attention机制的诈骗电话文本分类 [J ] . 小型微型计算机系统 , 2023 , 44 ( 11 ): 2502 - 2509 .

Zhou J J , Xu H K , Lu J K , et al . Position embedding and attention are introduced into the fraudulent phone text classification [J ] . Journal of Chinese Computer Systems , 2023 , 44 ( 11 ): 2502 - 2509 .

司海平 , 李阔 , 李婷婷 , 等 . 基于提示对比学习的小样本电信诈骗文本分类方法研究 [J ] . 计算机应用与软件 , 2025 : 1- 8 . ( 2025-01-27 ).

Si H P , Li K , Li T T , et al . Research on few-shot telecom fraud text classification approaches based on prompt-driven contrastive learning [J ] . Computer Applications and Software , 2025 : 1- 8 . ( 2025-01-27 ).

邓诗琦 , 洪亮 . 面向智能应用的领域本体构建研究: 以反电话诈骗领域为例 [J ] . 数据分析与知识发现 , 2019 , 3 ( 7 ): 73 - 84 .

Deng S Q , Hong L . Constructing domain ontology for intelligent applications: case study of anti tele-fraud [J ] . Data Analysis and Knowledge Discovery , 2019 , 3 ( 7 ): 73 - 84 .

刘卓娴 , 石拓 , 胡啸峰 . 电信网络诈骗警情要素抽取与模式挖掘研究 [J ] . 情报杂志 , 2025 , 44 ( 6 ): 168 - 176, 192 .

Liu Z X , Shi T , Hu X F . Research on alarm texts element extraction and pattern mining of telecom network fraud [J ] . Journal of Intelligence , 2025 , 44 ( 6 ): 168 - 176, 192 .

李衡峰 , 郑榕 , 邓菁 , 等 . 基于大模型的电信网络诈骗预警技术研究 [J ] . 警察技术 , 2025 ( 4 ): 13 - 18 .

Li H F , Zheng R , Deng J , et al . Research on telecom network fraud early warning technology based on large language model [J ] . Police Technology , 2025 ( 4 ): 13 - 18 .

Shen Z T , Wang K Z , Zhang Y Q , et al . Combating phone scams with LLM-based detection: where do we stand? (student abstract) [J ] . Proceedings of the AAAI Conference on Artificial Intelligence , 2025 , 39 ( 28 ): 29487 - 29489 .

Singh G , Singh P , Singh M . Advanced real-time fraud detection using RAG-based LLMs [PP ] . V1. arXiv ( 2025-01-25 )[ 2025-09-23 ] . arXiv: 2501.15290 .

Ma Z M , Wang P D , Huang M H , et al . TeleAntiFraud-28k: an audio-text slow-thinking dataset for telecom fraud detection [PP ] . V4. arXiv ( 2025-08-18 )[ 2025-09-23 ] . arXiv: 2503.24115 .

裴炳森 , 李欣 , 吴越 . 基于ChatGPT的电信诈骗案件类型影响力评估 [J ] . 计算机科学与探索 , 2023 , 17 ( 10 ): 2413 - 2425 .

Pei B S , Li X , Wu Y . Influence evaluation of telecom fraud case types based on ChatGPT [J ] . Journal of Frontiers of Computer Science and Technology , 2023 , 17 ( 10 ): 2413 - 2425 .

Guan S P , Cheng X Q , Bai L , et al . What is event knowledge graph: a survey [J ] . IEEE Transactions on Knowledge and Data Engineering , 2023 , 35 ( 7 ): 7569 - 7589 .

Li Z Y , Zhao S D , Ding X , et al . EEG: knowledge base for event evolutionary principles and patterns [M ] // Social Media Processing: Communications in Computer and Information Science . Singapore : Springer Singapore , 2017 : 40 - 52 .

Ding X , Li Z Y , Liu T , et al . ELG: an event logic graph [PP ] . V2. arXiv ( 2019-08-07 )[ 2025-09-23 ] . arXiv: 1907.08015 .

周胜利 , 徐睿 , 陈庭贵 , 等 . 生成式人工智能驱动下电信网络诈骗风险演化实证研究 [J ] . 电信科学 , 2025 , 41 ( 5 ): 149 - 165 .

Zhou S L , Xu R , Chen T G , et al . Empirical study on the evolution of telecom fraud risks driven by artificial intelligence generated content [J ] . Telecommunications Science , 2025 , 41 ( 5 ): 149 - 165 .

斯彬洲 , 孙海春 , 吴越 . 基于大语言模型和事件融合的电信诈骗事件风险分析 [J ] . 数据分析与知识发现 , 2025 , 9 ( 7 ): 38 - 51 .

Si B Z , Sun H C , Wu Y . Risk analysis of telecom fraud events based on large language models and event fusion [J ] . Data Analysis and Knowledge Discovery , 2025 , 9 ( 7 ): 38 - 51 .

Perera P , Oza P , Patel V M . One-class classification: a survey [PP ] . V1. arXiv ( 2021-01-08 )[ 2025-09-23 ] . arXiv: 2101.03064 .

Sun C J , Ji J , Shang B Y , et al . Overview of CCL23-Eval Task 6: telecom network fraud case classification [C ] // Proceedings of the 22nd Chinese National Conference on Computational Linguistics (Volume 3: Evaluations) . Harbin : Chinese Information Processing Society of China , 2023 : 193 - 200 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于冻结主干双回路进化的多智能体潜空间策略研究

企业AI智能体基础设施的层次化架构：基于云计算实践的分析框架

基于LLM-RFE-XGBoost方法的客户升级投诉预警研究

基于预训练大模型的业务流程实体与关系的自动提取和评估

基于通信网络大模型的智能运维人机交互系统研究与应用