摘要
加权建模是必要的,微分建模是重要的,把二者结合起来,进行加权微分建模既必要也重要.给出了常用模型的微分建模结果,讨论了加权建模中的计算和权重选择问题,探讨了加权微分建模的思路和方法,并结合典型数据验证了该方法的有效性和稳定性.象加权建模一样,加权微分建模的精度、实用价值等,是和权重确定得合理与否紧密相联;应先进行模拟,以与近期实际值或典型样本相差最小的参数所对应的模型为准.
Weighted modeling is necessary, differential modeling is important, combining them together, carrying through weighted differential modeling is both necessary and important. This paper presents the results of common models＇ differential modeling, discusses the issues when calculating and selecting the weights, probes into weighted differential modeling＇s reason and method, and validates the method＇s validity and stability. Like weighted modeling, weighted differential modeling＇s precision, practical value and so on, tie up closely with the weights＇ rationality; so we should carrying through simulating, select the model corresponding with the parameters which have the least differences with the factual value in the near future or representative samples.
出处
《数学的实践与认识》
CSCD
北大核心
2006年第9期1-7,共7页
Mathematics in Practice and Theory
基金
国家自然科学基金重点项目(90102013)成果
关键词
回归建模
加权建模
微分建模
加权微分建模
regressive modeling
weighted modeling
differential
weighted differential modeling
