电话诈骗识别系统的设计与实现  

Design and Implementation of Telephone Fraud Recognition System

作　　者：张杰俊 李爽[2] ZHANG Jie-jun;LI Shuang(China Telecom Corporation Limited Shanghai Branch,Shanghai 200040,China;Beijing University of Posts and Telecommunications,Beijing 100876,China)

机构地区：[1]中国电信上海分公司,上海200040 [2]北京邮电大学,北京100876

出　　处：《软件》2020年第4期190-194,共5页Software

摘　　要：本文将当前频发的电话诈骗问题与人工智能技术相结合,设计并实现了一个基于卷积神经网络的电话诈骗识别系统,该系统首先清洗并处理已采集的大批量用户话单数据,将其构建成用户特征图片,然后使用基于卷积神经网络技术构造的电话诈骗识别模型,对输入的用户通话特征图片进行挖掘,识别出诈骗人员的用户行为,从而帮助人们在通话初期提前识别出诈骗电话。Combining artificial intelligence with the frequent problems of telecommunication fraud,this paper designs and implements a telecommunication fraud recognition classifier based on convolutional neural network.After cleaning and processing the user’s telephone list data into pictures,the classifier constructed by convolutional neural network is used to mine the characteristics of fraudulent user's telephone data and identify the fraudulent user’s behavior.It can help people identify fraudulent users in advance and reduce potential property safety hazards.

关 键 词：电话诈骗 卷积神经网络 特征图构造 用户行为识别 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]