The goal of extendedFamily is to add new links to R’s generalized linear models. These families are drop in additions to the existing families.
You can install the released version of extendedFamily from CRAN with:
And the development version from GitHub with:
The heart data contains info on 4,483 heart attack victims. The goal is to predict if a patient died in the next 48 hours following a myocardial infarction.
library(dplyr, warn.conflicts = FALSE)
library(yardstick)
#> For binary classification, the first factor level is assumed to be the event.
#> Set the global option `yardstick.event_first` to `FALSE` to change this.
library(extendedFamily)
data(heart)
heart %>%
group_by(death) %>%
summarise(Count = n())
#> # A tibble: 2 x 2
#> death Count
#> <dbl> <int>
#> 1 0 4307
#> 2 1 176The low frequency of deaths suggests the loglog link is probably a better model than a logit link. Lets find out!
glmLogit <- glm(formula = death ~ anterior + hcabg + kk2 + kk3 + kk4 + age2 + age3 + age4,
data = heart, family = binomial(link = "logit"))
glmLoglog <- glm(formula = death ~ anterior + hcabg + kk2 + kk3 + kk4 + age2 + age3 + age4,
data = heart, family = binomialEF(link = "loglog"))Note the minimal code change between the two models. Only the family changed.
Lets calculate AUC.
predictions <- heart %>%
select(death) %>%
mutate(death = factor(death, levels = c("1", "0")),
logitProb = predict(object = glmLogit, newdata = heart, type = "response"),
loglogProb = predict(object = glmLoglog, newdata = heart, type = "response"))
roc_auc(data = predictions, truth = death, logitProb)
#> # A tibble: 1 x 3
#> .metric .estimator .estimate
#> <chr> <chr> <dbl>
#> 1 roc_auc binary 0.797
roc_auc(data = predictions, truth = death, loglogProb)
#> # A tibble: 1 x 3
#> .metric .estimator .estimate
#> <chr> <chr> <dbl>
#> 1 roc_auc binary 0.801A slightly higher AUC was achieved by simply changing the link.