支持向量机补偿的多因素灰色模型话务量预测

J4 ›› 2014, Vol. 36 ›› Issue (07): 1330-1335.

支持向量机补偿的多因素灰色模型话务量预测

郭勤1，贾振红1，覃锡忠1，盛磊2，陈丽2

(1.新疆大学信息科学与工程学院，新疆乌鲁木齐 830046；2.中国移动通信集团新疆有限公司，新疆乌鲁木齐 830063)

收稿日期:2013-03-28 修回日期:2013-05-31 出版日期:2014-07-25 发布日期:2014-07-25
基金资助:
中国移动通信集团新疆有限公司研究发展基金资助项目(XJM201201)

A multi-factor grey model based on SVM residual
error compensation for telephone traffic prediction

GUO Qin1,JIA Zhenhong1,QIN Xizhong1,SHENG Lei2,CHEN Li2

(1.School of Information Science and Engineering,Xinjiang University,Urumqi 830046;2.Subsidiary Company of China Mobile in Xinjiang,Urumqi 830063,China)

Received:2013-03-28 Revised:2013-05-31 Online:2014-07-25 Published:2014-07-25

摘要/Abstract

摘要：

为了提高对话务量的预测精度以及建模的速度，针对当前移动通信话务量预测受到多种因素的影响，提出了基于支持向量机残差补偿的多因素灰色话务量预测模型。该模型通过灰色关联分析法确定影响话务量的主因素变量，然后用多变量灰色模型进行预测，再用粒子群优化的最小二乘支持向量机进行残差序列预测，以实现残差补偿。实验结果表明，该预测模型具有所需样本小、预测精度高的优点，为话务量网络管理提供了一种新的预测工具。

关键词: 关联分析, 多因素灰色模型, 残差补偿, 话务量

Abstract:

In order to improve the prediction accuracy of telephone traffic and the speed of modeling, aiming at that the current mobile telephone traffic forecasting is influenced by many factors, a multifactor grey traffic prediction model based on Support Vector Machine (SVM) residual error compensation is proposed. In this model, the main factors affecting the traffic variables are determined by the grey correlation analysis method, and the multivariable grey model is used to do prediction. Then, the residual error sequence is predicted by the leastsquares SVM optimized by the particle swarm optimization so as to realize the residual error compensation. The experimental results show that the required samples of this prediction model are small, the model has the advantage of high accuracy, and the model provides a new forecasting tool for the telephone traffic network management.

Key words: correlation analysis;multifactor grey model;residual error compensation;traffic

郭勤1，贾振红1，覃锡忠1，盛磊2，陈丽2. 支持向量机补偿的多因素灰色模型话务量预测[J]. J4, 2014, 36(07): 1330-1335.

GUO Qin1,JIA Zhenhong1,QIN Xizhong1,SHENG Lei2,CHEN Li2. A multi-factor grey model based on SVM residual
error compensation for telephone traffic prediction [J]. J4, 2014, 36(07): 1330-1335.

参考文献

［1］Zhang Lei. The linear regression analysis of traffic forecast telecom revenue research［D］.Guangzhou:Sun Yatsen University, 2009.(in Chinese)
［2］Jiang Jianzhong. Time series analysis in the application of mobile telephone traffic forecasting ［D］. Beijing:Beijing University of Posts and Telecommunications, 2005.(in Chinese)
［3］Yu Yanhua,Wang Jun.Application of product seasonal ARIMA model of traffic forecast and result analysis［J］. Computer Engineering and Application, 2009,45(20):99102.(in Chinese)
［4］Liu Junbo, Sun Yan, Wang Daming, et al. A Kalman filterbased algorithm for realtime forecasting of communication traffic［J］. Journal of Communications Technology, 2012,45(6):8688.(in Chinese)
［5］Deng Bo,Li Jian, Sun Tao. Traffic prediction based on neural network ［J］. Journal of Chengdu Information Engineering College, 2008,23(5):521518.(in Chinese)
［6］Chen Dianbo, Xu Fucang, Wu Min. Traffic based on clustering and support vector machine forecasting model ［J］. Journal of Control Engineering, 2009,16(2):195198.(in Chinese)
［7］Subhrakanti N G. A dynamic quantization and power allocation for multisensor estimation of hidden Markov model［J］. IEEE Transactions on Automatic Control, 2012,57(7):16411656.
［8］Geng Liyan,Zhang Wei,Zhao Peng.LSSVM based on grey correlation analysis of railway freight volume prediction［J］. Journal of Railway, 2012,34(3):16.(in Chinese)
［9］Shen Jihong, Zhang Changbin, Li Jide. The prediction of ship motion via updating MGM(1,n) model［C］∥Proc of the IEEE International Conference on Grey Systems and Intelligent Services, 2009:533537.
［10］Chen Weimin, Chen zhigang. Network traffic prediction based on PSRLSSVM measurement ［J］. Journal of Computer Science, 2012,39(7):9295.(in Chinese)
［11］Wu ChihHung, WH S, YaWei Ho. A study on GPS GDOP approximation using supportvector machines［J］. IEEE Transactions on Instrumentation and Measurement, 2011,60(1):137145.
［12］Fu Lihua, Li Hongwei, Zhang Meng,et al. With multiple multiscale kernel functions, the radial basis function networks［J］. Journal of Huazhong University of Science and Technology (Natural Science Edition), 2010,38(1):3942.(in Chinese)
［13］Varshney S, Pandit L S M. Comparison of PSO models for optimal placement and sizing of statcom［C］∥Proc of Sustainable Energy and Intelligent System(SEISCON 2011), 2011:125133.
附中文参考文献：
［1］张蕾. 基于线性回归分析的话务预测电信收入的研究［D］. 广州：中山大学, 2009.
［2］蒋建忠. 时间序列分析在移动话务量预测中的应用［D］.北京：北京邮电大学,2005.
［3］于艳华，王军. 应用乘积季节ARIMA模型的话务量预测及结果分析［J］. 计算机工程与应用, 2009,45(20):99102.
［4］刘军搏，孙岩，王大鸣，等. 一种基于Kalman的实时话务量预测算法［J］. 通信技术, 2012,45(6):8688.
［5］邓波，李建，孙涛. 基于神经网络的话务量预测［J］.成都信息工程学院学报,2008,23(5):518521.
［6］陈电波，徐福仓，吴敏. 基于聚类和支持向量机的话务量预测模型［J］. 控制工程, 2009,16(2):195198.
［8］耿立艳，张伟，赵鹏. 基于灰色关联分析的LSSVM铁路货运量预测［J］. 铁道学报, 2012,34(3):16.
［10］陈卫民，陈志刚. 基于PSRLSSVM的网络流量预测［J］. 计算机科学, 2012,39(7):9295.
［12］付丽华，李宏伟，张猛,等. 带多个核函数的多尺度径向基函数网络［J］. 华中科技大学学报（自然科学版）, 2010,38(1):3942.

[1]	雷轩, 程光, 张玉健, 郭靓, 张付存. 基于电力网络态势感知平台的告警信息关联分析[J]. 计算机工程与科学, 2023, 45(07): 1197-1208.
[2]	赵国生, 王甜甜, 王健. 一种边缘设备动态信任度的评估模型[J]. 计算机工程与科学, 2021, 43(09): 1574-1583.
[3]	李杰，楼芳，金渝筌，董智馨. 面向APT攻击的关联分析检测模型研究[J]. J4, 2015, 37(08): 1458-1464.
[4]	金涛1,2，牛亚峰2. 一种复杂产品界面可用性定量评估方法的研究和应用[J]. J4, 2015, 37(04): 819-823.
[5]	曹卫东,刘红霞. 基于多层次灰色关联分析的复杂网络节点排序模型[J]. J4, 2014, 36(06): 1165-1171.
[6]	郭军, 赵和鹏, 朱长仁. 一种基于自适应权值灰关联分析的毫米波/红外融合识别方法[J]. J4, 2010, 32(2): 146-149.
[7]	韩宗芬杨志玲储杰涂旭平. 一种用于网络安全系统的报警聚类与关联模型[J]. J4, 2005, 27(10): 8-9.