uscale.RdExtract the (normalized) utility scale for an elimination-by-aspects (EBA) model.
uscale(object, norm = "sum", log = FALSE)an object of class eba, typically the result of a
call to eba
either sum, so the scale values sum to unity (default);
or a number from 1 to number of stimuli, so this scale value becomes one;
or NULL (no normalization)
should the log of the utility scale values be returned? Defaults to FALSE.
Each utility scale value is defined as the sum of aspect values (EBA model parameters) that characterize a given stimulus. First these sums are computed for all stimuli, then normalization (if any) is applied. As each type of normalization corresponds to a multiplication by a positive real number, the ratio between scale values remains constant.
The (normalized) utility scale of the stimuli.
data(drugrisk)
A <- list(c(1), c(2,7), c(3,7), c(4,7,8), c(5,7,8), c(6,7,8))
eba1 <- eba(drugrisk[, , group = "male30"], A) # EBA model
uscale(eba1) # sum-to-unity normalization
#> alc tob can ecs her coc
#> 0.04264735 0.02178031 0.02385802 0.17877250 0.53993276 0.19300905
uscale(eba1, norm=1) # u(alcohol) := 1
#> alc tob can ecs her coc
#> 1.0000000 0.5107072 0.5594256 4.1918779 12.6604047 4.5256982
uscale(eba1, norm=5) # u(heroine) := 1
#> alc tob can ecs her coc
#> 0.07898642 0.04033893 0.04418702 0.33110142 1.00000000 0.35746869
uscale(eba1, norm=NULL) # no normalization
#> alc tob can ecs her coc
#> 0.04027701 0.02056976 0.02253199 0.16883632 0.50992327 0.18228160
uscale(eba1, norm=1, log=TRUE) # log utility scale, log u(alcohol) := 0
#> alc tob can ecs her coc
#> 0.0000000 -0.6719589 -0.5808448 1.4331488 2.5384794 1.5097719