基于C#与R语言的重复测量设计资料方差分析的自动化实现

Automatic implementation of ANOVA of repeated measurement design data based on C#and R language

目的基于C#语言和R语言开发一款便于临床研究人员使用的软件,用于实现重复测量设计资料方差分析的自动化。方法首先用C#语言将统计分析数据导入到软件中;然后调用R语言命令脚本,完成重复测量设计资料的方差分析与结果输出;最后通过2020年出版的第5版《医学统计学》教材中的实例,结合实例结果和SPSS操作结果,来验证本自动化实现软件运行的效果。结果统计分析结果包括了基本统计描述、重复测量方差分析与组内两两比较的结果图表。基于C#语言和R语言相结合开发出的统计软件运行结果与教材手工计算分析、SPSS操作的结果完全一致。其操作简单、结果直观,在效率上优于手工分析和SPSS菜单点选的反复操作;同时最大程度地避免了手工分析和SPSS操作步骤中的低级错误;且结果自动化呈现,减少了结果判断与选择的错误,确保了结果的准确性和规范性。结论该软件能够自动实现重复测量设计资料的单、双因素两/多水平方差分析,值得在临床上推广使用。

Objective Based on C#language and R language,a software is developed,which is convenient for clinicians to use.It is used to realize the automation of variance analysis of repeated measurement design data.Methods First,the statistical analysis data are imported into the software by C#language,then the R language command script is invoked to complete the ANOVA and the result output of the repeated measurement design data.Finally,the example in the 5th edition of medical statistics published in 2020 is used,combined with the example results and SPSS operation results,to verify the effect of the automation of the software.Results The results of statistical analysis included basic statistical description,repeated measurement ANOVA and result chart of pairwise comparison within the group.Based on the combination of C#language and R language,the running results of the statistical software are completely consistent with the results of manual calculation and analysis of teaching materials and SPSS operation.The operation is simple and the results are intuitive,which greatly surpasses the repeated operation of manual analysis and SPSS menu selection in efficiency;at the same time,the low-level errors in manual analysis and SPSS operation steps are avoided to the greatest extent,and the results are presented automatically,which reduces the errors in judgment and selection of results,and ensures the accuracy and standardization of results.Conclusion The software can automatically realize the single factor,two factor and two/multi-level ANOVA of repeated measurement design data,which is worthy of clinical application.