基于通信大数据挖掘的网络业务类型划分方法研究  

Research on the Method of Network Service Classification Based on Communication Big Data Mining

作　　者：李强[1] 金雨超[1] 郭景赞[1] 

机构地区：[1]中国联合网络通信有限公司网络技术研究院,北京100044

出　　处：《移动通信》2017年第3期51-54,共4页Mobile Communications

摘　　要：在传统的Gn数据解析过程中,业务类型里"未知服务"的占比通常在70%以上。为了提高解析结果的分析价值,结合数据分析和数据挖掘技术对DNS解析流量字段进行细分,该设计方案在最大化利用原始数据的同时,能从多维度细分用户网络业务,优化结果可以为用户画像、用户标签、用户群体特征等分析应用提供有力支持。In the process of the traditional Gn data parsing, the proportion of ＂unknown＂ services is always above 70%. In order to improve the value of parsing result, the DNS traffic field was divided in detail according to data analysis and data mining techniques. Maximizing the utilization of the original data, the proposed solution decomposes users＇ network services in multi-dimensions of which optimization results can provide powerful support to analysis applications of the user portrait, user tag and user group characteristics.

关 键 词：数据挖掘 数据解析 网络业务类型划分 聚类算法 

分 类 号：TN915[电子电信—通信与信息系统]