摘要
用户的投诉行为既受到客观网络环境作用,又具有强烈的主观行为意识。通过收集投诉用户的主客观数据特征,引入机器学习算法预测潜在的投诉用户,提早介入进行网络优化和用户关怀,能够有效降低网络投诉的发生。本文介绍了一种基于XGBoost算法的投诉用户特征识别和预测方法,通过收集投诉用户数据特征,构建投诉特征指标集,利用XGBoost算法迭代处理实现潜在投诉用户预测。经大量数据验证,本算法可有效实现对网络潜在投诉预测,为网络优化提供了重要依据。
Users'complaint behaviours are aff ected by both the objective network environment and subjective behaviours.By collecting the subjective and objective data characteristics of complaining users,introducing machine learning algorithms to predict potential complaining users,and intervening early to carry out network optimisation and user care,the occurrence of network complaints can be effectively reduced.This paper introduced a complaint user feature identification and prediction method based on the XGBoost algorithm.By collecting complaint user data features,constructing a complaint feature index set and using the iterative processing of the XGBoost algorithm to achieve potential complaint user prediction.The algorithm had been verified by the practical implementation of the system,which effectively achieved the potential complaint prediction for the whole network and could provide an important basis for network optimization.
作者
宗宇雷
梁童
吴伟嘉
阙鋆淑
ZONG Yu-lei;LIANG Tong;WU Wei-jia;QUE Yun-shu(China Mobile Group Design Institute Co.,Ltd.,Beijing 100080,China;China Mobile Group Yunnan Co.,Ltd.,Kunming 650000,China)
出处
《电信工程技术与标准化》
2023年第11期7-12,共6页
Telecom Engineering Technics and Standardization
