单变量回归确定移动平均权重算法在话务量预测中的应用研究  

The research of the application in teletraffic prediction using weighted moving average method based on univariate regression

作　　者：严宇平 吴广财 江疆 

机构地区：[1]广东电网有限责任公司信息中心,广东广州510620

出　　处：《电子设计工程》2015年第22期44-46,49,共4页Electronic Design Engineering

摘　　要：话务量是客服中心运营管理的重要方面,呼叫中心的高效率运作是电网系统客户满意度以及大规模停电预警及突发状况处理的重要保证,而坐席人员排班又是呼叫中心运作的重要问题。科学合理的坐席安排能更好地提高呼叫中心的运作效率,不仅降低了呼叫中心的运营成本,从而提升电力呼叫中心的服务水平。本文提出在移动平均法中引入加权,利用预测期与历史各期的单变量回归,将系数作归一化处理作为权重,进行话务量的预测。文章以真实场景中的实验结果证明了本文所提方法的有效性。Teletraffic is an important part in operation management of the call center.Forecasting of the teletraffic is an effective way to improve service levels of the call center. In this paper, we propose to use weighted moving average for teletraffic forecasting. The coefficients are calculated by univariate regression on data from predictive and historical period, and then these coefficients are normalized for teletracffic forecasting as weights of weighted moving average method. Experiments on real world data show the effectiveness of our method.

关 键 词：话务量预测 单变量回归 加权移动平均 权重计算法 

分 类 号：TN99[电子电信—信号与信息处理]