摘要
数据组合处理方法(GMDH)是20世纪70年代发展起来的一种启发式自组织建立模型的方法；它能充分地、合理地利用数据,自动进行变量组合,筛选以及判断从而得到合适的模型；简单介绍了该方法建模的基本原理和算法实现,给出了仿真算例,并与用相同资料建立的PPR预测模型的预测结果进行了比较；仿真结果表明,用GMDH方法建立非线性系统模型,具有预测精度高、计算稳定性好等优点。
Group method of data handling （GMDH） is a method of modeling by heuristic self--organization in 1970＇s. It can use data abundantly, combine variables by itself and find the model by self --selection and comparison. A fundamental idea and implement algorithm of the method are simply introduced. A numerical simulation example is given. The comparisons of prediction results between the GMDH model and the PPR model are also performed using the same data. Simulation result shows that GMDH method can be applied to nonlinear system modeling successfully and has such advantages as higher prediction precision and good calculation stability, etc.
出处
《计算机测量与控制》
CSCD
2006年第7期939-941,共3页
Computer Measurement &Control
基金
NSAF联合基金(10276005)
国防行业重点预研项目(421010401-3)。
关键词
数据组合处理方法
非线性系统
建模
预测
group method of data handling
nonlinear system
modeling
prediction
