基于大数据技术的疑似骚扰和电信欺诈电话监测系统设计与实现

Design and Implementation of Bothering and Fraud Phone Call(BFPC) Monitoring System Based on Big Data Technology

大数据挖掘一般基于Hadoop/Spark平台,这些开源项目在许多方面做的很好,但也存在一些不足,如缺乏高性能No SQL数据库支持和需要复杂的手工编程。文章介绍一种基于第二代大数据技术,通过类SQL智能大数据语言实现疑似骚扰和电信欺诈电话监测系统,通过分析通话特征,设计疑似骚扰和电信欺诈通话识别算法模型,利用大数据技术对话单数据、网络数据、信令数据等多种数据源进行大数据分析和处理,结合网内和网间的客户投诉信息,能够有效识别疑似骚扰和电信欺诈通话。

Hadoop and Spark are the common big data platforms for data mining.These open source projects do perform very well in many cases, but a few problems exist, including lacking high performance NoSQL database support and requiring manual programming. This paper presents a design and implementation of a Bothering and Fraud Phone Call （BFPC） Monitoring System based on a Second Generation Big Data platform. The entire system is implemented without manual programming. Everything is done through a SQL-like language （JIMO SQL）. The system analyzes the characteristics of BFPC based on phone call patterns and defnes BFPC recognition algorithms in JIMO SQL. It uses big data technologies to sift through a variety of data sources including Call Detail Records （CDRs）, Network Traffc Data, and Phone Signaling Data, and combines with customer complaint information, so as to effectively identify BFPCs.