r - 通过连续变量预测分类变量

Question

我正在进行逻辑回归分析，以通过多个连续变量预测分类变量（性别）。但是，我遇到了一些错误，真的需要有人帮助我。

这里我放了模拟数据和代码。

library(tidyverse)
library(broom)
library(purrr)

data <- data.frame(cat = rep(c("A", "B", "C", "D"), each = 20),
                      shar = rep(c("fr", "bs"), each = 20),
                      sex = rep(c("f", "m"), each = 20),
                      bmi = runif(20, min = 0, max = 30),
                      fbg = runif(20, min = 70, max = 140),
                      AB = runif(20, min = 0, max = 1),
                      Ac = runif(20, min = 0, max = 1),
                      Ad = runif(20, min = 0, max = 3),
                      Ae = runif(20, min = 0, max = 4),
                      Af = runif(20, min = 2, max = 4),
                      Ba = runif(20, min = 2, max = 10),
                      Bb = runif(20, min = 10, max = 40),
                      Bc = runif(20, min = 5, max = 50),
                      BD = runif(20, min = 0.5, max = 1),
                      BF = runif(20, min = 5, max = 10),
                      Ca = runif(20, min = 2, max = 10),
                      Cb = runif(20, min = 10, max = 40),
                      Cc = runif(20, min = 5, max = 50),
                      CD = runif(20, min = 0.5, max = 1),
                      CF = runif(20, min = 5, max = 10))

data1 <- data %>% 
  gather(Fac, value, -cat:-sex) %>% 
  split(.$Fac)  %>% 
  map(~glm(sex ~ value,family = binomial(link = "logit"), data =.) %>%  
        tidy(.) %>% 
        select(estimate, std.error,statistic,p.value) %>% 
        slice(2)) %>% 
  bind_rows(.id = "variables") 

这是错误

eval(family$initialize) 中的错误：y 值必须为 0 <= y <= 1。

然后，我修改了代码，但仍然出现错误

data$gendr <- ifelse(data$sex =="m",1L,0L)
data1 <- data %>% 
  gather(Fac, value, -cat:-sex) %>% 
  split(.$Fac)  %>% 
  map(~glm(gendr ~ value,family = binomial(link = "logit"), data =.) %>%  
        tidy(.) %>% 
        select(estimate, std.error,statistic,p.value) %>% 
        slice(2)) %>%  ## Extracting regression of genes
  bind_rows(.id = "varaibles")

错误是

eval 中的错误（predvars、data、env）：找不到对象“gendr”

谢谢！

标签： rglm