大型呼叫中心话务量预测  

Forecast on traffic of large call center

作　　者：牟颖[1] 王俊峰[1] 谢传柳[1] 夏正洪[2] 

机构地区：[1]四川大学计算机学院,四川成都610064 [2]中国民航飞行学院空中交通管理学院,四川广汉618307

出　　处：《计算机工程与设计》2010年第21期4686-4689,4719,共5页Computer Engineering and Design

基　　金：国家863高技术研究发展计划基金项目(2008AA01Z208;2009AA01Z405);四川省青年基金项目(2009-28-419)

摘　　要：为进行呼叫中心的坐席数估计和后续的排班工作,分析了历史话务量数据的特点,总结出影响大型呼叫中心话务量的因素,并用这些影响因素的不同组合来预测话务量,通过结果的对比分析得出相对最优的话务量预测模型。在此模型的基础上分别采用BP神经网络算法和支持向量机算法(LS-SVM)对话务量进行了预测,通过分析和比较结果表明,BP神经网络比支持向量机算法更适合对大型呼叫中心话务量的预测。To estimate the number of agents and the follow-up scheduling of a call center, the characteristics of traffic＇ s historical data are analyzed. Some factors impacting the traffic of large call center are summed up, and different combinations of these factors are used to predict the traffic, compared the results, a optimum model is obtained. Based on the model, the BP neural network algorithm and LS-SVM algorithm are used to the traffic forecast, besides a comparative analysis is made for the two methods. According to the presented simulations, it is proved that BP neural network algorithm is more efficient to the traffic forecasting for large call center than LS-SVM.

关 键 词：呼叫中心 坐席数 话务量 BP神经网络 支持向量机算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]