电信科学 ›› 2019, Vol. 35 ›› Issue (5): 17-24.doi: 10.11959/j.issn.1000-0801.2019099

• 专题:智能通信技术及应用 • 上一篇    下一篇

移动通信网络投诉热点问题智能预警方法

朱琳,赵娟,王伊婷,冯俊兰,邓超   

  1. 中国移动通信有限公司研究院,北京 100053
  • 修回日期:2019-05-09 出版日期:2019-05-20 发布日期:2019-05-21
  • 作者简介:朱琳(1983- ),女,博士,中国移动通信有限公司研究院人工智能算法高级技术研究员,主要研究方向为网络智能化运营。|赵娟(1993- ),女,现就职于中国移动通信有限公司研究院,主要研究方向为机器学习。|王伊婷(1993- ),女,现就职于中国移动通信有限公司研究院,主要研究方向为机器学习。|冯俊兰(1974- ),女,博士,中国移动通信有限公司研究院人工智能和智慧运营研发中心总经理、首席科学家,主要从事大数据挖掘和人际交互领域的研究工作。|邓超(1978- ),男,博士,中国移动通信有限公司研究院人工智能与智慧运营研发中心总经理助理、高级工程师,AIIA联盟专家委员会委员,主要从事人工智能、机器学习、大规模并行数据挖掘相关研究工作。

Smart prediction of the complaint hotspot problem in mobile network

Lin ZHU,Juan ZHAO,Yiting WANG,Junlan FENG,Gchao DEN   

  1. China Mobile Research Institute,Beijing 100053,China
  • Revised:2019-05-09 Online:2019-05-20 Published:2019-05-21

摘要:

在移动通信网络中,一旦发生投诉热点问题,通常会影响大量用户的上网、通话体验,进而引发大量用户投诉,即网络投诉热点问题,该问题因其影响范围大、用户多,一旦发生往往影响恶劣,需实时监控并提前预警和处理。提出一种基于用户级信令数据的投诉热点问题预警方法。首先基于对业务逻辑的理解,选择了S1接口数据中与用户体验相关的30个关键字段;然后,提取one-hot特征、统计衍生特征和差分特征3类特征来详细刻画用户感知状况;针对数据中噪声大及正负样本不均衡等问题,采用泛化能力较强且针对样本不均衡问题有所改善的 LightGBM 分类器实时识别受影响的用户。这一方法可以实时输出潜在受影响的用户与区域,先于用户投诉提前处理问题或进行客户关怀,有效降低影响,提升客户体验。试验结果与某省现网部署应用均验证了算法的有效性。

关键词: 投诉热点问题智能预警, 信令数据, 特征提取, LightGBM分类器

Abstract:

In telecom communication network,a hot customer complaint often affects hundreds even thousands of users’ service and leads to significant economic losses and bulk complaints.An approach was proposed to predict a customer complaint based on real-time user signaling data.Through analyzing the network business layer logic,30 key segments related to the user experience in the S1 interface data were selected.Further,one-hot features,statistical derived features,and differential features were extracted to classify user perceptions in detail.Considering the problems of noise data and unbalanced training samples,LightGBM was chosen to train the prediction model.Experiments are conducted to prove the effectiveness and efficiency of this proposal.As of today,this approach has been deployed in our daily business to locate the hot complaint problem scope as well as to report affected users and area.

Key words: prediction of the complaint hotspot problem, signaling data, feature extraction, LightGBM classifier

中图分类号: 

No Suggested Reading articles found!