摘要
提出了一种基于混沌理论和支持向量的预测方法。通过重构相空间的饱和嵌入维数,确定支持向量机的最佳输入变量;通过计算混沌序列的最大Lyapunov指数,确定支持向量机预测模型的最大有效预测步数;利用支持向量机强大非线性映射能力、网络结构的自动最优化特性,实现时间序列的非线性预测。最后,应用于某型发动机压气机的试车时间序列数据建模与分析,结果证明该方法具有较高的预测精度。
A new support vector forecasting model based on chaos theory is presented in this paper. It adopts support vector machines as nonlinear forecaster and determines network＇s input variable number through computing reconstruct phase space＇s saturated embedding dimension. The maximum effective forecasting steps is determined by computing chaos time series＇ largest Lyapunov exponent. It makes use of support vector machines＇ strongly nonlinear mapping ability, and network＇s structure is optimally auto- created. Application results in aeroengine show that the presented method possesses much better precision.
出处
《火力与指挥控制》
CSCD
北大核心
2007年第7期54-57,共4页
Fire Control & Command Control
关键词
相空间重构
支持向量机
航空发动机
李雅普诺夫指数
预测
reconstruct phase space, support vector machines, aeroengine, Lyapunov exponent, forecasting
