Do linear and logistic regression on R take range?-CodePudding

I've encountered challenges when using a glm and lm in R that they don't accept range of variables.

For example my code is this

model1 = glm(RESPONSE~., train.df, family= "binomial")
summary(model1)

RESPNSE is a variable name on the dataset. I am aware that the dot"." means all variables.

Is there away that I can tell R to take a range of variables, for example from 1 to N? N = integer numbers

The reason asking this question is because I need to include many, it takes a lot of time to type all of them with no mistakes.

Your help is greatly appreciated.

CodePudding user response：

Using the built in anscombe data frame this regresses y1 on the range from x3 to x4. The variable names can be replaced with column numbers or even a mix of variable names and column numbers.

lm(y1 ~ ., subset(anscombe, select = c(y1, x3:x4)))

lm(y1 ~ ., subset(anscombe, select = c(5, 3:4)))

We can alternately express this using pipes:

anscombe |>
  subset(select = c(y1, x3:x4)) |>
  lm(y1 ~ ., data = _)

Using dplyr it is nearly the same:

library(dplyr)
anscombe %>%
  select(y1, x3:x4) %>%
  lm(y1 ~ ., data = .)