Graphical Research On Sample Size Calculation Of Linear Regression And Logistic Statistical Model

Background:The linear regression and logistic regression model is widely used in various types of medical research,but for the specific needs of its corresponding sample size,no relevant research has been elaborated for a long time.More often than not,some of the general guidelines summarized by statisticians based on experience in long-term research work were applied.Since the forties and fifties of the last century,statisticians have begun to pay attention to the sample size estimation method of multiple regression models.Therefore,researches on the method of estimating the sample size of the linear regression and logistic regression model have begun to appear gradually at home and abroad.After entering the 20 th century,relevant researches have gradually increased,providing a rich theoretical material for this research.At the same time,as the related researches have been developing rapidly,new problems are coming one after another.Most formulas of multivariate regression models calculating the sample size are complex,with a lot of parameters and large calculations,and the values of some parameters are not continuous,which undoubtedly hinders the promotion and application of related methods.Objectives:To conclude the progress of sample size estimation methods for linear regression and logistic regression models and simplify the knowledge structure of related methods.Meanwhile,construct a graphical sample size calculation platform for human-computer interaction,avoiding complex sample size calculation formula and make its software implementation simpler and more reliable.Methods:Through the systematic retrieval and literature combing of the linear regression and logistic regression models sample size methodology,we summarized the sample size calculation methods of linear regression and logistic regression,and selected sample estimation methods which included common conditions of common models,the widest application scope,and simple parameters easy to understand.With the help of some software such as shiny and ggplot2 in the R software,we created a graphical formula to calculate the web page.In the design of the graphical interface,we fully considered the theoretical value range of the selected formula parameters,the calculation formulas for different models in different conditions and the different values of the parameters.In different situations,the corresponding graphical interface of the sample size calculation formula was applied to calculate the sample size.Results:The main results of this study are as follows:(1)concluded the development process of linear regression and logistic regression sample size calculation methods,and summarized the knowledge structure of sample size calculation methods for different cases.(2)completed the interface design of the graphical calculation of linear regression and logistic regression sample size calculation method under different conditions;(3)realized the graphical calculation of linear regression and logistic regression sample size calculation method under different conditions(4)the calculation interface provided in this paper can reasonably limit the range of parameter values in the sample size calculation formula to avoid errors caused by unreasonable parameter values.Conclusions:The graphical sample size calculation interface constructed in this paper has a good practical value as it can better calculate the sample size of the linear regression and logistic regression models,and simplify the calculation process of the relevant sample size.