CODEWITHSUNDEEP
rregression
Adrien
Adrien

Reputation: 151

Best way to save regression equations in R ?

I have a large number of regression equations that I would like to save in R and I am not sure how to do it efficiently. For example, 

y1 ~ x1 + x2 + x3 + x4  (country A) 
y1 ~ x1 + x2 +      x4  (country B) 
y1 ~ x1 + x2 + x3 + x4  (country C) 
y1 ~         + x3 + x4  (country D)

Ideally, I would like to be able to answers such as how many times x2 occurred ? 3. what is the most common variable ? x4

Should I save everything in a list ? or is there a better method ?

Upvotes: 2

Views: 125

Answers (2)

duffymo
duffymo

Reputation: 308938

Wouldn't you prefer a model like this and glm?

y ~ x1 + x2 + x3 + x4 + country

Let the regression tell you which factors drop out.

Upvotes: 0

Thomas
Thomas

Reputation: 44555

Put them in a list:

myformulas <- 
list(a = y1 ~ x1 + x2 + x3 + x4,
     b = y1 ~ x1 + x2 +      x4,
     c = y1 ~ x1 + x2 + x3 + x4,
     d = y1 ~         + x3 + x4)

You can then perform operations on them like:

# what variables are in which formulae
> str(lapply(myformulas, function(x) attr(terms(x), 'term.labels')))
List of 4
 $ a: chr [1:4] "x1" "x2" "x3" "x4"
 $ b: chr [1:3] "x1" "x2" "x4"
 $ c: chr [1:4] "x1" "x2" "x3" "x4"
 $ d: chr [1:2] "x3" "x4"

# where is `x1` used?
> str(lapply(myformulas, function(x) 'x1' %in% attr(terms(x), 'term.labels')))
List of 4
 $ a: logi TRUE
 $ b: logi TRUE
 $ c: logi TRUE
 $ d: logi FALSE

# how many times is each variable used?
> table(unlist(lapply(myformulas, function(x) attr(terms(x), 'term.labels'))))
x1 x2 x3 x4 
 3  3  3  4

From that structure you could easily answer your questions about the use of different variables.

Upvotes: 4

Related Questions