摘要
目的 介绍医用非线性统计分析模型及其NoSA实现。方法 建立各种非线性模型及模型参数的最大似然估计方法 ,并附有实例演算。结果 由于NoSA嵌入了网络算法技术 ,解释变量可以是定量、定性和有序的 ,并能分析变量间的交互作用可对变量作逐步选择。结论 NoSA的非线性模型模块计算稳定 ,并能自动排除广义共线关系对模型参数最大似然估计值的影响。此模块较SPSS有其明显的优点。
Objective To introduce modular of NoSA for nonlinear regression analysis.Methods a dosen nonliear regression models and its maximum likelihood functions were discribed and some classical examples were analyzed.Results Since an advanced net algorithm was built in NoSA,the explanatory varibales can be quantitative,qualitative and ordinal,and interactions among explanatory regressors could be included in models,and selection of regressors could be realised by employing stepwise algorithm.Conclusion NoSA can be used to deal with data in ill condition,the generalized multicollinearities can be eliminated automatically.Considering model options and advantage of parameter estimation algorithm,NoSA is better than SPSS.
出处
《中国卫生统计》
CSCD
北大核心
2000年第5期279-281,共3页
Chinese Journal of Health Statistics
基金
国家自然科学基金!资助项目 (编号 :3960 0 1 2 8 39770 677)
关键词
非线性模型
最大似然估计
数理统计
统计分析
Nonliear model Maximum likelihood estimation Mathematical statistics