基于逐步回归分析的话务预测研究  

Research on telephone traffic forecasting based on stepwise regression analysis

作　　者：卢盛祺[1,2] 邵俊[2] 韩景倜[1] 

机构地区：[1]上海财经大学信息管理与工程学院,上海200433 [2]复旦大学软件学院,上海200433

出　　处：《微型机与应用》2015年第1期88-92,共5页Microcomputer & Its Applications

基　　金：国家自然科学基金(71271126);教育部博士点专项科研基金(20120078110002)

摘　　要：针对现有话务预测模型变量单一、预测误差大等问题引入了逐步回归分析法;结合海河银行信用卡呼叫中心实际需求,对话务数据进行了抽取集成等预处理操作;随之建立了各影响因素的一元回归模型,并在此基础上得到了多元逐步回归的话务预测模型;最后将此建模方法与其他分析法进行了综合比较并给出评价。Firstly, the method of stepwise regression analysis was introduced, aiming at the chief problem that existing in current traffic forecasting model. And combined with actual demand of the credit card call center of Haihe bank, the pre-processing of the traffic data such as extraction and integration were done. Then the unitary regression models were constructed. And on this basis, traffic forecasting model of multivariate stepwise regression was studied in details. Finally, this method was contrasted with other analysis method and overall evaluations were made.

关 键 词：话务预测 回归分析 逐步回归 时间序列分析 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]