摘要
采用提升小波方法构造出一种满足双正交的小波函数,并将这种小波函数作为支持向量机的核函数;此外,用线性规划问题来代替二次规划问题及稀疏正则化,本质上确保了解的稀疏性。基于提升小波构造出提升小波支持向量机模型,并将其用于交通流量的预测中。仿真实验表明该模型具有良好的预测能力和泛化能力。
This paper presented a lifting wavelet to construct a new wavelet function which could be used as an allowable kernel function for support vector machine （SVM）. Proposed a linear programming algorithm to replace quadratic programming in support vector machine solution process and sparse regularization, and guaranteed the sparseness of the solution. Based on the lifting wavelet, constructed and used a SVM model in the prediction of traffic flow. The simulation results illustrate the proposed method has well ability for prediction and generalization.
出处
《计算机应用研究》
CSCD
北大核心
2007年第8期275-277,共3页
Application Research of Computers
基金
四川省教育厅重点基金资助项目(0229957)
中国教育部博士点培养基金资助项目(20040613013)
关键词
提升小波
支持向量机
交通流量预测
lifting wavelet
support vector machine
traffic flow predictive
