基于人工智能算法的投诉用户质差小区定位及预测的方法研究

详细信息    查看全文 | 推荐本文 |

英文篇名：Research on the analysis method of location and prediction of complaint user subdistrict based on artificial intelligence algorithm 作者：刘丽娟 英文作者：LIU Li-juan;China Mobile Group Guangdong Co., Ltd.Foshan Branch; 关键词：逻辑回归 ; 决策树 ; 质差定位 ; 质差预测 英文关键词：logistic regression;;decision tree;;differential localization;;low quality prediction 中文刊名：电信工程技术与标准化 英文刊名：Telecom Engineering Technics and Standardization 机构：中国移动通信集团广东有限公司佛山分公司; 出版日期：2019-05-15 出版单位：电信工程技术与标准化 年：2019 期：05 语种：中文; 页：22-27 页数：6 CN：11-4017/TN ISSN：1008-5599 分类号：TN929.5;TP18

摘要

为了解决自动定位质差小区,本文创新性提出以数据驱动来实现业务规则自动化,充分考虑用户投诉的时间和问题、网络交互信令、无线性能指标和告警等海量数据,运用人工智能技术中的逻辑回归及决策树算法得出2G语音、VoLTE语音及数据业务投诉质差小区定界模型、质差小区预测模型及问题网元,更快速、更智能、更准确地输出语音及数据业务投诉原因、质差小区及质差原因,并对潜在质差小区进行预测,对无线规划、基站建设、故障处理、网络改造等方面的工作也具有积极指导意义。
        In order to solve the automatic positioning complaints and inferior community, innovative offered to data driven automation to implement business rules, give full consideration to the user to complain about the time and signaling, network interaction, wireless performance indicators and alarms, such as huge amounts of data, using artificial intelligence technology in the logistic regression and gradient descent algorithm for the 2 G voice、VoLTE and data services complaints and inferior, poor quality, and the village community and bound model and the problem of network, more quickly, more intelligent, more accurate output reasons of voice and data service complaints reasons, poor quality, and the community and bad reasons, and to forecast the potential bad district, it also has a positive guiding signifi cance for wireless planning, base station construction, fault treatment, network transformation and other aspects of work.

引文