线性模型循环

Question

我正在对各种基线特征如何影响并发症进行单变量分析。以下是我观察个体特征SEX、BMI和AGE的片段，以及它们如何单独影响COMPLICATION。

我想做的是让它也能循环通过等并发症

COMPLICATION1
COMPLICATION2
COMPLICATION3

这是我的样本代码

#Generate list of predictors for univariate logistic model
varlist <- c(“SEX",
“BMI”,
”AGE")
#Run univariate logistic model based on predictors using "COMPLICATION" as an outcome
lapply(varlist, function(x) {
mod <- glm(reformulate(x, 'COMPLICATION'), 
data = NSQIP_D.clean, 
family = binomial)
summary.glm(mod)$coefficients
}) -> results
#Combine Model results into one file
results <- do.call(rbind, results)
#Write table of results
write.table(results, file="Univariate_Results.txt", sep = "t", row.names = TRUE)

Answer 1

你的意思是这样的吗：

require("tidyverse")
lapply(c("COMPLICATION1", "COMPLICATION2", "COMPLICATION3"), function(y) {
lapply(c("SEX", "BMI", "AGE"), function(x) {
glm(paste0(y, " ~ ", x), data = NSQIP_D.clean, family = binomial) %>% 
summary() %>% 
.$coefficients %>% 
return() 
}) %>% 
do.call(rbind, .) %>% 
cbind(y, .) %>% 
return()
}) %>% 
do.call(rbind, .) %>% 
write.table(file = "Univariate_Results.txt", sep = "t", row.names = T)

在这里，您为每个复杂情况运行三个单变量回归，并将结果存储在联合文本文件中。我用不同的语法写了这篇文章。然而，你也可以用基本的R表示。