运营商客户满意度模型构建

The model construction of user tariffs and network satisfaction of telecommunications operating companies

基础电信运营企业的客户满意度,外部测评采用抽样问卷访谈的方法。该方法受样本数的影响通常会有一定偏差。通过用户与运营商交互的行为数据来建立模型,可以实现内部满意度测评,以便更好把握客户感知。基于用户交互数据的满意度模型构建包括关联分析,权重分析,预测感知三个主要环节。在实施层面,数据汇聚重点针对用户全旅程触点数据和用户特征数据,数据清洗涉及缺失值处理,无量纲化;建模算法重点解决快速识别和判断负向感知客户,通过决策树自动分箱+机器学习算法+否决指标的方式构建模型。

The sampling questionnaire interview method used for external evaluation of user tariffs and network satisfaction of ba⁃sic telecommunications operating companies usually has certain deviations due to the number of samples.However,a model is estab⁃lished based on the behavioral data of interaction between users and operators to achieve internal satisfaction.Measure and evaluate to better understand customer perceptions.The construction of satisfaction model based on user interaction data includes three main links:correlation analysis,weight analysis,and predictive perception.At the implementation level,data aggregation focuses on the us⁃er's full journey touch point data and user characteristic data.Data cleaning involves missing value processing and dimensionality;the modeling algorithm focuses on quickly identifying and judging customers with derogatory satisfaction,and automatically binning through decision trees+Machine learning algorithm+veto indicator method to build the model.