摘要
对农药残留预测模型可靠性问题进行了分析.根据矩阵的条件数理论,提供了检验预测模型可靠性的方法,基于岭回归和广义岭回归估计理论,建立了对不可靠预测模型改良的方法.最后通过氰戊菊酯在甘篮上的残留动态预测,对所建改良方法进行了检验,结果表明。
According to the theory of condition number of matrix, this paper provided a method for testing the reliability of predicting model,and on the basis of ridge regression theory,the method for improving unreliable predicting model was built, which was tested by predicting the residue dynamics of fenvalerate in cabbages.The results showed the precision of predicting model was increased by a big range.
出处
《应用生态学报》
CAS
CSCD
1999年第5期599-602,共4页
Chinese Journal of Applied Ecology
基金
吉林省科委资助项目
关键词
农药残留
预测模型
可靠性
检验方法
改良
Model for predicting insecticide residue,Reliability,Testing method,Improvement method.