基于支持向量机的多因素话务量预测研究  

Research on multi factor traffic forecasting based on support vector machine

作　　者：曾雨桐 钱学荣[1] 

机构地区：[1]南京邮电大学通信与信息工程学院,江苏南京210003

出　　处：《微型机与应用》2016年第1期63-66,共4页Microcomputer & Its Applications

摘　　要：提高移动通信话务量的预测精度对提高网络性能、增进用户体验具有重要意义。由于多种因素会影响到移动通信话务量的准确预测,故选择多因素灰色话务量预测模型来预测话务量。先对数据进行预处理,用关联分析法找到影响话务量预测的主要因素。但此模型对波动较大的数据预测精度较低,用支持向量机的模型来改善预测结果,选取拥有较强的敛散性和全局寻优能力的复高斯小波核函数优化向量机。从仿真结果可以看出该模型有更好的收敛作用和较为理想的预测效果。It is very important to improve the network performance and user experience by forecasting mobile communication traffic vol- ume accurately. According to the characteristics of traffic load, the multi factor grey model for traffic forecast is put forward. It needs to find main factors which affect the traffic volume forecast by using correlation analysis method firstly. Due to the low prediction accu- racy of data fluctuation, so it selects complex Gaussian wavelet kernel function which has strong convergence and divergence and global optimization ability to optimize the vector machine. The simulation results show that the model has better convergence effect and pre- diction results.

关 键 词：复高斯小波核函数 支持向量机 多因素 话务量预测 

分 类 号：TN929.5[电子电信—通信与信息系统]