I want to repeat a statistical function (like -lm-, -glm- or -glmgee-) for a lot of variables. But, it does not work for statistical functions (example 1) but works for simple functions (example 2).
Important: I do not mean multivariate regression and using cbind()!
Example 1:
a = rnorm(10, 5, 1)
b = rnorm(10, 7, 1)
c = rnorm(10, 9, 1)
d = rnorm(10, 10, 1)
i = list(a, b, c)
for (x in i) {
lm(x~d)
}
Example 2:
a = rnorm(10, 5, 1)
b = rnorm(10, 7, 1)
c = rnorm(10, 9, 1)
d = rnorm(10, 10, 1)
i = list(a, b, c)
for (x in i) {
plot(x+d)
}
You can check in this site: https://rdrr.io/snippets/
You need to save the output in an object. In the case of lm(), the best is to save the results in a list. This is achieved automatically if you'd use lapply() instead of for().
models
Thanks Jochen Wilhelm
It worked.
Is there a post-estimation function to address p-value and coefficient of each regression?
Something like coef(model) or pval(model) .... with addressing each covariate
You get all that via the summary() method of lm. This is a list containing (among other things) a coefficients matrix with a row for each coefficient (including the intercept) and the columns "Estimate", "Std. Error", "t value", and "Pr(>|t|)". To extract all the estimates of the second coefficients (the first is the intercept) from the fitted models you may thus call
sapply(models, function(model) summary(model)$coefficients[2 , "Estimate"])
I use sapply() here instead of lapply() as this simplifies the result to a numeric vector. If you use lapply() you will get a list returned.
Thanks a lot Jochen Wilhelm
I know summary()
I was looking for post-estimation addressing
I found it in your kind answer as:
sapply(models, function(model) summary(model)$coefficients[2 , "Estimate"])
What is the equivalent of $coefficients for "Pr(>|t|)"?
Thanks again
- Badges
- Science topic
- Similar topics
- Correction