基于相关分析的一元非线性回归模型优选方法探讨  

Study on Optimal Selection Method of the Unitary Nonlinear Regression Model Based on Correlation Analysis

作　　者：周忠军[1] 张尊沛[1] 易杰忠[2] 陈恩会[2] 张浩[1] 

机构地区：[1]沛县农林局,江苏沛县221600 [2]徐州市农业局,江苏徐州221006

出　　处：《中国农学通报》2008年第9期132-136,共5页Chinese Agricultural Science Bulletin

摘　　要：就生物科学农业系统中的一元非线性回归模型问题,在相关分析的基础上,按照"残差最小"原则,应用"残差"数据序列,采取关联分析、相对误差、均方差比值、小误差概率等方法进行综合评判,优选模型精度高,与实际相符,且避免了错误应用模型事件发生,为实践应用提供了科学评判依据和手段。In this paper, a optimal selection method is presented to solve the problem of the unitary nonlinear regression model in bioscience agriculture system. According to the principle of residual error error probability. The precision of the evaluation was high and the result of the evaluation corresponded to reality. Using the optimal selection method could avoid the mistake of model application. The author provided scientific basis and means for practice and application.

关 键 词：非线性关系 非线性模型 优选法 综合评判 

分 类 号：S117[农业科学—农业基础科学]