CODEWITHSUNDEEP
rlistglm
appleLover
appleLover

Reputation: 15691

Make List of GLM in R

I have a bunch of models I want to use through GLM, so I want to put them in a list, and loop through the list. However, the following codes draws an error:

m1<- glm(target~total+tot_eit_h_h1+tot_both_h_h1, data=data, family='binomial')
m2<- glm(target~total+tot_both_h_h1, data=data, family='binomial')
models<- c(m1,m2)

This works perfectly:

predictions <- predict(m1, data, type='response')

This code draws an error:

predictions <- predict(models[1], data, type='response')

Error in UseMethod("predict") : 
no applicable method for 'predict' applied to an object of class "list"

Upvotes: 3

Views: 2666

Answers (2)

Metrics
Metrics

Reputation: 15458

You can use Map function with the trick as follows (example illustrated with mtcars data):

dep<-list("mpg~","mpg~")  # list of dep variables which is same in both models
indep<-list("cyl","cyl+disp") #for model 1 includes cyl and for model 2 includes cyl and disp
prediction<-Map(function(x,y) predict(lm(as.formula(paste(x,y)),data=mtcars)),dep,indep)

> prediction[[1]]
          Mazda RX4       Mazda RX4 Wag          Datsun 710      Hornet 4 Drive   Hornet Sportabout 
           20.62984            20.62984            26.38142            20.62984            14.87826 
            Valiant          Duster 360           Merc 240D            Merc 230            Merc 280 
           20.62984            14.87826            26.38142            26.38142            20.62984 
          Merc 280C          Merc 450SE          Merc 450SL         Merc 450SLC  Cadillac Fleetwood 
           20.62984            14.87826            14.87826            14.87826            14.87826 
Lincoln Continental   Chrysler Imperial            Fiat 128         Honda Civic      Toyota Corolla 
           14.87826            14.87826            26.38142            26.38142            26.38142 
      Toyota Corona    Dodge Challenger         AMC Javelin          Camaro Z28    Pontiac Firebird 
           26.38142            14.87826            14.87826            14.87826            14.87826 
          Fiat X1-9       Porsche 914-2        Lotus Europa      Ford Pantera L        Ferrari Dino 
           26.38142            26.38142            26.38142            14.87826            20.62984 
      Maserati Bora          Volvo 142E 
           14.87826            26.38142

Upvotes: 3

Drew Steen
Drew Steen

Reputation: 16607

Try

models<- list(m1,m2)

and

predictions <- predict(models[[1]], data, type='response')

glm returns a list. The c operator 'flattens' that list into a vector, so that the functions associated with glm objects won't work on the flattened vector. As a rule, you don't want to use c to concatenate lists (even if you can sometimes get away with it, as in your example).

Upvotes: 4

Related Questions