如何从 R 中的回归 model 列表中提取对象？

Question

我想编写一个 function my_function ，它将 dataframe 作为输入。 在这个 function 中，我运行了一个逻辑回归来预测影响观察是否被禁止的决定的重要因素。 显着性定义为 p 值小于 0.05 ( p-value < 0.05 )。 我不知道如何提取这些因素。

my_function(data){
   fit <- glm(is_prohibited ~ ., data, family = "binomial")
}

我不知道如何检查重要性的因素并提取我需要的因素，因为我只能提取 model 的系数。

数据：

structure(list(is_prohibited = c("No", "No", "No", "No", "No", 
"No", "No", "No", "No", "No", "No", "No", "No", "No", "No", "Yes", 
"Yes", "Yes", "Yes", "Yes", "Yes", "Yes", "Yes", "Yes", "Yes", 
"Yes", "Yes", "Yes", "Yes", "Yes"), weight = c(69L, 79L, 82L, 
81L, 84L, 81L, 64L, 76L, 77L, 88L, 86L, 84L, 89L, 84L, 86L, 82L, 
84L, 88L, 88L, 89L, 86L, 76L, 87L, 67L, 70L, 76L, 71L, 91L, 72L, 
88L), length = c(53L, 52L, 54L, 50L, 48L, 51L, 53L, 52L, 53L, 
52L, 46L, 52L, 52L, 50L, 47L, 54L, 54L, 50L, 49L, 50L, 50L, 57L, 
47L, 50L, 50L, 51L, 52L, 52L, 48L, 54L), width = c(17L, 21L, 
20L, 23L, 19L, 20L, 16L, 20L, 23L, 23L, 19L, 17L, 22L, 23L, 23L, 
24L, 20L, 21L, 20L, 17L, 20L, 18L, 21L, 24L, 21L, 23L, 18L, 21L, 
17L, 20L), type = c("Suitcase", "Bag", "Suitcase", "Bag", "Suitcase", 
"Bag", "Suitcase", "Bag", "Suitcase", "Bag", "Suitcase", "Bag", 
"Suitcase", "Bag", "Suitcase", "Bag", "Suitcase", "Bag", "Suitcase", 
"Bag", "Suitcase", "Bag", "Suitcase", "Bag", "Suitcase", "Bag", 
"Suitcase", "Bag", "Suitcase", "Bag")), row.names = c(NA, 30L
), class = "data.frame")

Answer 1

我相信您正在寻找 broom::tidy:

fit <- glm(as.factor(is_prohibited) ~ ., data, family = "binomial")

library(broom)
tidy(fit)

  term         estimate std.error statistic p.value
  <chr>           <dbl>     <dbl>     <dbl>   <dbl>
1 (Intercept)  -0.798     11.3      -0.0704   0.944
2 weight        0.00713    0.0534    0.133    0.894
3 length        0.0201     0.166     0.121    0.903
4 width        -0.0330     0.174    -0.189    0.850
5 typeSuitcase -0.265      0.822    -0.322    0.747

tidy(fit) 返回一个标准数据框，您可以在其中访问系数、p 值等以进行进一步计算。