均匀设计试验多元回归分析方法在新型生物农药研制中的应用  

Multiple-regression Analysis Method of Uniform Design Experiment Applied to the Development of New Biological Pesticide

作　　者：刘宇锋[1] 苏天明[1] 苏利荣[1] 秦芳[1] 李琴[1] 梁芷姮 

机构地区：[1]广西农业科学院农业资源与环境研究所,南宁530007

出　　处：《农药》2018年第2期105-110,共6页Agrochemicals

基　　金：广西自然科学基金项目(2016GXNSFCA380008);广西科学研究与技术开发计划项目(桂科合15104001-25);南宁市邕宁区科学研究与技术开发计划项目(20170101C)

摘　　要：[目的]均匀设计作为一种考虑试验点"均匀散布"原则为基础的科学试验设计方法,获得广泛应用。但均匀设计试验数据进行多元回归分析的具体方法与步骤,回归模型分析结果比较与选择,优化方案获得与数据验证,相关方法与步骤还比较模糊,是限制均匀设计方法推广应用的一个技术瓶颈。[方法]结合新型生物农药优化配方研制的具体实例,重点介绍在试验因素与试验结果间回归模型关系不明确的情况下,利用SPSS回归分析和Excel规划求解功能进行均匀设计数据的多元回归分析的具体步骤。[结果]明确了优化方案选择和试验结果验证的方法,形成均匀设计数据分析的基本流程。[结论]为均匀设计在新型生物农药配方研制提供了一种现实的数据分析解决方案。[Aims] Uniform design experiment based on ＇uniform distribution＇ of experimental sites, as a scientific design method, has been widely used. However, methods and steps of multiple regression analysis for uniform design experimental data, comparison and selection of regression analysis models, optimization scheme acquisition and data validation, and correlation methods and steps remain unsolved, which are the technical bottleneck restricting the application ofuniforrn design experiment. [Methods] In this study, taking an example of the formulation optimization of a new biological pesticide, the procedures of multiple-regression analysis for the uniform design experimental data were obtained using regression analysis model of SPSS software and programming solution module of Microsoft Excel when the relationships between the experimental variables and results were unclear. [Results] The methods of optimization scheme selection and experimental result verification were determined, and the steps of multiple- regression analysis for uniform design experimental data were obtained. [Conclusions] The results can provide a practical data analysis method for the uniform design experiment applied to the formulation development of new biological pesticides.

关 键 词：均匀设计 多元回归分析 SPSS 生物农药 

分 类 号：TQ450[化学工程—农药化工]