Logistic regression double loop R

Question

I would like to run a logistic regression on each variable pair in my dataset, excluding the pairs that have already been regressed. All variables are binomial. The output should include the pair tested and test statistics. Since I have different datasets I have to use it on, I am trying to write a script that will work on different datasets containing various number of variables that are all binomial.

Sample dataset contains 6 variables named Var1:Var6 with 50 observations each.

Var1 = c(rbinom(50,1,0.5))
Var2 = c(rbinom(50,1,0.25))
Var3 = c(rbinom(50,1,0.6))
Var4 = c(rbinom(50,1,0.2))
Var5 = c(rbinom(50,1,0.3))
Var6 = c(rbinom(50,1,0.8))

dt = data.table(Var1, Var2, Var3, Var4, Var5, Var6)  
head(dt)

  Var1 Var2 Var3 Var4 Var5 Var6
1    1    0    1    1    0    1
2    1    0    0    0    0    1
3    0    0    1    0    0    1
4    1    0    1    0    1    0
5    1    0    1    1    0    1
6    0    1    1    1    0    0

So I would like to regress Var1 on Var2:Var6, Var2 on Var3:Var6 etc. The output table should contain Dependent_var, Independent_var, Estimate, Stat, P_value.

I've made an output table:

n = ncol(dt)    
output <- data.table(matrix(nrow=(n*(n+1))/2, ncol=5))
names(output) = c("Dependent_var", "Independent_var", "Estimate", "Stat", "P_value")
head(output)

 Dependent_var Independent_var Estimate Stat P_value
1:            NA              NA       NA          NA      NA
2:            NA              NA       NA          NA      NA
3:            NA              NA       NA          NA      NA
4:            NA              NA       NA          NA      NA
5:            NA              NA       NA          NA      NA
6:            NA              NA       NA          NA      NA

Now I am not sure how to loop one variable with all and then loop over all of them... AND then fill up the output table correctly... Any help is very much appreciated!

Samuel · Accepted Answer · 2017-10-18 19:25:48Z

2

I have sketched out a simple procedure, see if it helps:

my_func <- function(x) {
  for (i in 1:ncol(x)) {
    fit <- glm(x[, i] ~ ., data = x, family = binomial(link = "logit"))
    print(summary(fit))
  }
}

The parameter x is a dataframe.

edited Oct 18, 2017 at 19:25

answered Oct 18, 2017 at 19:10

Samuel

3,0695 gold badges32 silver badges50 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Panda Over a year ago

Thanks for your quick answer @jsb!

Panda Over a year ago

I seem to be doing something wrong though... So I turned my dt into a dataframe and then tried to run the loop. Am not getting any data though, how can I tickle it out? p.s. sorry, I can't seem to get the code properly formatted in the comments section.

dt <- data.frame(dt)  my_func <- function(dt) {       for (i in 1:ncol(dt)) {              fit <- glm(dt[, i] ~ ., data = dt, family = binomial(link = "logit"))              print(summary(fit)) }  }

Collectives™ on Stack Overflow

Logistic regression double loop R

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related