摘 要：“呼叫中心”一词来源于英文 Call Center，自其诞生以来就作为企业和客户之间的沟通桥梁，是客户和企业沟通最直接的渠道，因此优化呼叫中心的接听应答效率是管理者一直以来的追求。该文为了向管理者提供前瞻的话务量预测信息供管理者决策，提出了使用 XGBoost、LightGBM、Catboost 算法结合信息价值分析法选择的特征通过累加型滑动窗口法建立话务量预测模型，并在真实数据上比较了三个算法的预测表现。结果表明 XGBoost 算法对于运营商呼叫中心话务量的预测较为准确，为坐席排班提供数据支撑。
关键词：呼叫中心；XGBoost；话务量；坐席排班
DOI:10.19850/j.cnki.2096-4706.2021.22.025
中图分类号：TP18 文献标识码：A 文章编号：2096-4706（2021）22-0086-04
Call Center Seat Telephone-Traffic Volume Prediction Based on XGBoost Algorithm
ZHAO Long, ZHOU Yuan, LI Fei, FAN Wenbin
(Kedaduochuang Cloud Network Technology Co., Ltd, Hefei 230000, China)
Abstract: The word“call center”comes from English Call Center, since its birth, it has been used as a communication bridge between enterprises and customers, it is the most direct channel for customers and enterprises to communicate. Therefore, optimizing the answering efficiency of call center has always been the pursuit of managers. To provide managers with forward-looking forecast information of telephonetraffic volume for their decision-making, this paper proposes to use XGBoost, LightGBM and Catboost algorithms and combined with the characteristics selected by the information value analysis method to establish the telephone-traffic volume prediction model through the cumulative sliding window method, and compares the prediction performance of the three algorithms on the real data. The results show that the XGBoost algorithm is more accurate in predicting the call center’s telephone-traffic volume, and provide data support for seat scheduling.
Keywords: call center; XGBoost; telephone-traffic volume; seat scheduling
参考文献：
[1] BUIST E，CHAN W，L’ECUYER P. Speeding up call center simulation and optimization by Markov chain uniformization [C]//2008 Winter Simulation Conference.Miami：IEEE，2008：1652- 1660.
[2] AKTEKIN T，SOYER R. Call center arrival modeling：A Bayesian state space approach [J].Naval Research Logistics（NRL）， 2011，58（1）：28-42.
[3] CHASSIOTI E，WORTHINGTON D J. A new model for call centre queue management [J].The Journal of the Operational Research Society 2004，55（12）：1352-1357.
[4] CHEN T，GUESTRIN C. Xgboost：A scalable tree boosting system [C]//Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining.New York：Association for Computing Machinery，2016：785-794.
[5] KE G，MENG Q，FINLEY T，et al. Lightgbm：A highly efficient gradient boosting decision tree [J].Advances in neural information processing systems，2017，30：3146-3154.
[6] RANKA S，SINGH V. CLOUDS：A decision tree classifier for large datasets [C]//Proceedings of the 4th knowledge discovery and data mining conference.Syracuse University，1998：1-34.
[7] LI P，WU Q，BURGES C. Mcrank：Learning to rank using multiple classification and gradient boosting [J].Advances in neural information processing systems，2007，20：897-904.
[8] Microsoft-Corporation. Latest Document of LightGBM [EB/OL]. [2021-08-27].https://lightgbm.readthedocs.io/en/latest/Features.html.
[9] SHI H. Best-first decision tree learning [D].Hamilton：The University of Waikato，2007.
[10] PROKHORENKOVA L，GUSEV G，VOROBEV A，et al. CatBoost：unbiased boosting with categorical features [J/OL].arXiv： 1706.09516 [cs.LG].（2017-06-28）.https://arxiv.org/abs/1706.09516v4.
[11] KOHAVI R，LI C H. Oblivious decision trees，graphs， and top-down pruning [C]//Fourteenth IJCAI.Montreal：IJCAI，1995： 1071-1079.
作者简介：赵龙（1982—），男，汉族，安徽铜陵人，副总裁， 硕士，研究方向：通信运营商IT咨询规划、软件系统设计、智慧社区、 云计算和数据智能；周源（1991—），男，汉族，安徽合肥人， 算法工程师，硕士，研究方向：数据挖掘和自然语言处理；李飞 （1982—），男，汉族，安徽利辛人，总经理，硕士，主要研究方向： 通信运营商 IT 咨询规划、软件系统设计、大数据平台建设、数据 建模和数据智能；范文斌（1990—），男 ，汉族，安徽黄山人， 部门经理，本科，研究方向：软件系统设计、数据智能、知识图谱。