J4 ›› 2011, Vol. 29 ›› Issue (02): 97-.

• 论文 • 上一篇    下一篇

基于神经网络的客服中心话务量预测模型

张一农1,刘伯龙2,王文婷1   

  1. 1.吉林大学 通信工程学院,长春 130012;2.南京邮电大学 通信与信息工程学院,南京 210046
  • 出版日期:2011-03-25 发布日期:2011-04-25
  • 通讯作者: 刘伯龙(1988— ),男,长春人,南京大学本科生,主要从事神经网络及其在通信技术中的应用研究,(Tel)86-15996318883 E-mail:glliug@jlu.edu.cn。
  • 作者简介:张一农(1958— )|男|长春人|吉林大学工程师|主要从事无线通信网研究|(Tel)86-13843057059(E-mail)Zyn@jlu.edu.cn; 通讯作者:刘伯龙(1988— ),男,长春人|南京大学本科生|主要从事神经网络及其在通信技术中的应用研究,(Tel)86-15996318883 (E-mail)glliug@jlu.edu.cn。

Neural Network Based Traffic Prediction Model of Customer Service Center

ZHANG Yi-nong1|LIU Bo-long2,WANG Wen-ting1   

  1. 1. College of Communication Engineering, Jilin University,Changchun 130012,China;2.College of Telecommunications & Information Engineering, Nanjing University Posts and Telecommunications,Nanjing 210046,China
  • Online:2011-03-25 Published:2011-04-25

摘要:

针对现有预测模型在话务量发展趋势变化、新技术新业务引入后模型失效、预测精度下降等问题,提出一种基于神经网络和事件样本库的智能预测方法。该方法具有自学习功能,可根据预测误差自动调整预测参数并更新事件样本,对话务量趋势变化、事件影响程度变化及新事件的发生具有持续自适应能力。仿真结果表明,该预测方法能有效降低预测误差,与现有方法相比,话务量的预测精度提高了6.57%\.

关键词: 话务量, 神经网络, 预测模型

Abstract:

An intelligent prediction model based on neural network and event sample database is proposed to solve the changes of traffic trends and the model failure after the introduction of new technologies and new business and decrease of prediction accuracy. The method has self\|learning function and continuing adaptive capacity on changes of traffic trends, events influence and the occurrence of new events,which can automatically adjust predictive parameters and update the event samples according to predictive error. Simulation results show that the method effectively improves the accuracy of traffic prediction, also has a high value for investment planning of the company's customer service center.

Key words: traffic, neural network, prediction model

中图分类号: 

  • TP183