Tidymodels Logistic Regression getting coefficients and standard errors

Question

Is there a way to get the standard errors and p-values for logistic regression in tidy models?

I can get the coefficients by the following code below.. but I want to calculate odds ratios for each feature and I will need the standard errors as well..

glm.fit <- 
  logistic_reg(mode = "classification") %>%
  set_engine(engine = "glm") %>% 
  fit(Species ~ ., data = iris)


glm.fit$fit$coefficients

Usually you can do this by calling summary() on a glm object, but I'm trying to use tidymodels here.

score 4 · Answer 1 · answered Jul 12 '20 at 18:50

You can try:

library(broom)
library(tidymodels)

glm.fit <- 
  logistic_reg(mode = "classification") %>%
  set_engine(engine = "glm") %>% 
  fit(Species ~ ., data = iris)

tidy(glm.fit)

# A tibble: 5 x 5
  term         estimate std.error  statistic p.value
  <chr>           <dbl>     <dbl>      <dbl>   <dbl>
1 (Intercept)     16.9    457457.  0.0000370    1.00
2 Sepal.Length   -11.8    130504. -0.0000901    1.00
3 Sepal.Width     -7.84    59415. -0.000132     1.00
4 Petal.Length    20.1    107725.  0.000186     1.00
5 Petal.Width     21.6    154351.  0.000140     1.00

Tidymodels Logistic Regression getting coefficients and standard errors

1 Answers1