摘要
随着物流快递等业务的迅速发展,联络中心作为一个新兴服务行业也随之变成了服务机构和客户沟通的最重要的桥梁。联络中心任务量预测的准确性对基础设施和人员投入至关重要。因此,文中提出了一种结合小波变换和PSO-BP的组合预测模型,通过小波变换把任务量序列分解成高频和低频序列,再为分解序列建立合适的PSO-BP预测模型,求出最优解。最后,实例分析表明,该模型对非线性时间序列有更好的拟合能力和更高的预测精度。
With the rapid development of logistic,express delivery and so on,the contact center,as a new service industry,has become the most important bridge between the service organizations and their customers. The accuracy of tasks prediction in contact centers of logistic is very important to infrastructure investment and staffing. Therefore,a model that combines the wavelet transform and PSO- BP neural network is proposed. By the wavelet transform,the tasks are decomposed into high frequency and low frequency series,for which the suitable PSO- BP models are established to search the optimal solution. Finally,the analysis of the example indicates that the fitting ability and prediction accuracy of the method are better than other methods.
出处
《物流工程与管理》
2015年第9期114-115,109,共3页
Logistics Engineering and Management
基金
国家自然科学基金项目(编号:71572113
71303157
71102070
11171221
71271138
71202065
71103199
71371140)
上海市一流学科项目(编号:XTKX2012
S1201YLXK)
上海市教委创新基金(No.14YZ088
14YZ089)
上海市研究生和大学生创新项目(编号:151025213
SH2015056
SH2014054
SH2014062
XJ2015083)
关键词
联络中心
小波变换
粒子群算法
BP神经网络
任务量预测
contact center
wavelet transform
particle swarm optimization
back propagation(BP) neural network
tasks prediction