《電子技術(shù)應(yīng)用》
您所在的位置:首頁 > 通信與網(wǎng)絡(luò) > 設(shè)計應(yīng)用 > 一種基于圖神經(jīng)網(wǎng)絡(luò)的電信詐騙識別方法
一種基于圖神經(jīng)網(wǎng)絡(luò)的電信詐騙識別方法
2021年電子技術(shù)應(yīng)用第6期
張杰俊1,唐穎淳1,,季述鄖2,,李靜林2
1.中國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)國家重點(diǎn)實(shí)驗(yàn)室,,北京100876
摘要: 通信技術(shù)的普及給人們帶來便捷的同時,電信欺詐行為也急劇增加,。由于詐騙行為特征,、號碼類型等與正常業(yè)務(wù)具有極高相似性,傳統(tǒng)基于統(tǒng)計的電信欺詐檢測方法難于篩選,。提出將用戶通信關(guān)系轉(zhuǎn)換為一組拓?fù)涮卣?,建立通信社交有向圖,將具有統(tǒng)計特征的頂點(diǎn)表示用戶,,具有關(guān)系特征的邊表示他們之間的活動,。在通信社交圖基礎(chǔ)上,通過圖卷積模塊捕獲用戶的通信行為規(guī)律和通信社交關(guān)系特征,,通過池化讀出機(jī)制聚合通信社交網(wǎng)絡(luò)的潛在特征,,以識別電信欺詐行為。真實(shí)通信歷史數(shù)據(jù)驗(yàn)證表明了該方法的有效性,。
中圖分類號: TP18,;F626
文獻(xiàn)標(biāo)識碼: 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,,電信欺詐檢測成為了一個十分困難的問題,。

    傳統(tǒng)電信欺詐檢測技術(shù)主要基于用戶屬性和通話記錄來獲得用戶行為樣本,再通過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間的行為統(tǒng)計進(jìn)行分類,往往會出現(xiàn)時間尺度特征不足的問題,。同時,,由于用戶通話行為的復(fù)雜性,以固定窗口的統(tǒng)計特征作為詐騙電話的統(tǒng)計依據(jù)[3-4],,容易受到長期行為變化影響,,分類效果差。




本文詳細(xì)內(nèi)容請下載:http://forexkbc.com/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。