摘要
阐述了稀疏贝叶斯方法在时间序列预测中应用的理论基础,将稀疏贝叶斯方法应用于Log istic方程产生的混沌时间序列和发动机油滑数据的预测,并与支持向量机(SVM)和RBF神经网络时间序列预测进行了比较.实验结果表明,稀疏贝叶斯方法不仅具有SVM的性能,而且比SVM使用更少的核函数,取得了较好的预测效果.
The basic theoretic analysis of sparse Bayesian method in time series forecasting is introduced. Chaotic time series produced by Logistic equation and some type of engine lubrication time series are used for feasibility validation. In order to show its superiority, support vector machine （SVM） and RBF neural networks forecaster are also used during numerical simulations. Examples show that sparse Bayesian classification achieves comparable recognition accuracy to the SVM, and also requires substantially fewer kernel functions. Experimental results show the better performance in forecasting.
出处
《控制与决策》
EI
CSCD
北大核心
2006年第5期585-588,共4页
Control and Decision
基金
国家自然科学基金项目(60175011
60375011)
中国科技大学科学研究发展基金项目(030501F)
关键词
稀疏贝叶斯
支持向量机
非线性预测
RBF神经网络
Sparse Bayesian classification
Support vector machine
Nonlinear forecasting
RBF neural network
