《電子技術(shù)應(yīng)用》
您所在的位置:首頁(yè) > 通信与网络 > 设计应用 > 一种基于图神经网络的电信诈骗识别方法
一种基于图神经网络的电信诈骗识别方法
2021年电子技术应用第6期
张杰俊1,唐颖淳1,季述郧2,李静林2
1.中国电信股份有限公司上海分公司,上海200041; 2.北京邮电大学 网络与交换技术国家重点实验室,北京100876
摘要: 通信技术的普及给人们带来便捷的同时,电信欺诈行为也急剧增加。由于诈骗行为特征、号码类型等与正常业务具有极高相似性,传统基于统计的电信欺诈检测方法难于筛选。提出将用户通信关系转换为一组拓扑特征,建立通信社交有向图,将具有统计特征的顶点表示用户,具有关系特征的边表示他们之间的活动。在通信社交图基础上,通过图卷积模块捕获用户的通信行为规律和通信社交关系特征,通过池化读出机制聚合通信社交网络的潜在特征,以识别电信欺诈行为。真实通信历史数据验证表明了该方法的有效性。
中圖分類(lèi)號(hào): TP18;F626
文獻(xiàn)標(biāo)識(shí)碼: A
DOI:10.16157/j.issn.0258-7998.200976
中文引用格式: 張杰俊,唐穎淳,季述鄖,等. 一種基于圖神經(jīng)網(wǎng)絡(luò)的電信詐騙識(shí)別方法[J].電子技術(shù)應(yīng)用,2021,47(6):25-29,34.
英文引用格式: Zhang Jiejun,Tang Yingchun,Ji Shuyun,et al. A telecom fraud identification method based on graph neural net-
work[J]. Application of Electronic Technique,2021,47(6):25-29,34.
A telecom fraud identification method based on graph neural network
Zhang Jiejun1,Tang Yingchun1,Ji Shuyun2,Li Jinglin2
1.China Telecom Corporation Limited Shanghai Branch,Shanghai 200041,China; 2.State Key Laboratory of Networking and Switching Technology,Beijing University of Posts and Telecommunications, Beijing 100876,China
Abstract: While communication technology brings convenience to people, telecom fraud also increases sharply. Traditional detection methods are mainly based on data mining and statistical learning of history data. However, due to the high similarity between fraud behavior and normal business, traditional statistical methods are difficult to screen. This paper proposes to transform user communication relationship into a set of topological features and establish communication social directed graph, where vertices with statistical characteristics represent users and edges with relational characteristics represent activities between them. On the basis of the communication social graph, the potential characteristics of the communication social network are learned through the graph neural network, and the information characteristics of multiple nodes are aggregated through pooling readout mechanism, in order to identify the telecom fraud users. The validation of real communication history data shows the effectiveness of this method.
Key words : fraud detection;communication social network;graph neural networks;behavior classification

0 引言

    隨著信息社會(huì)的發(fā)展,電信欺詐高發(fā),但由于通信關(guān)系的復(fù)雜性和不確定性,電信欺詐檢測(cè)成為了一個(gè)十分困難的問(wèn)題。

    傳統(tǒng)電信欺詐檢測(cè)技術(shù)主要基于用戶(hù)屬性和通話(huà)記錄來(lái)獲得用戶(hù)行為樣本,再通過(guò)SVM、LGB等機(jī)器學(xué)習(xí)方法學(xué)習(xí)行為特征[1-2]。這些方法主要使用短時(shí)間的行為統(tǒng)計(jì)進(jìn)行分類(lèi),往往會(huì)出現(xiàn)時(shí)間尺度特征不足的問(wèn)題。同時(shí),由于用戶(hù)通話(huà)行為的復(fù)雜性,以固定窗口的統(tǒng)計(jì)特征作為詐騙電話(huà)的統(tǒng)計(jì)依據(jù)[3-4],容易受到長(zhǎng)期行為變化影響,分類(lèi)效果差。




本文詳細(xì)內(nèi)容請(qǐng)下載:http://www.ihrv.cn/resource/share/2000003568。




作者信息:

張杰俊1,唐穎淳1,季述鄖2,李靜林2

(1.中國(guó)電信股份有限公司上海分公司,上海200041;

2.北京郵電大學(xué) 網(wǎng)絡(luò)與交換技術(shù)國(guó)家重點(diǎn)實(shí)驗(yàn)室,北京100876)





wd.jpg

此內(nèi)容為AET網(wǎng)站原創(chuàng),未經(jīng)授權(quán)禁止轉(zhuǎn)載。

相關(guān)內(nèi)容