Plot ROC curves from an aba model — aba_plot

This function plots ROC curves across group - outcome - stat combinations and currently supports stat_glm.

Usage

aba_plot_roc(model, drop_facet_labels = FALSE)

Arguments

model: abaModel. A fitted aba model with stat_glm type.
drop_facet_labels: logical. Whether to remove facet labels from plots. The facet labels tell you what the group and outcome is for the plot. Sometimes these labels are unnecessary when you have only one group and one outcome, or when you want to add labels in another way.

Value

a ggplot with roc curves for all predictor sets across each group - outcome - stat combination

Examples

data <- adnimerge %>% dplyr::filter(VISCODE == 'bl')

# fit glm model with binary outcome variables
model <- data %>% aba_model() %>%
  set_groups(everyone()) %>%
  set_outcomes(ConvertedToAlzheimers, CSF_ABETA_STATUS_bl) %>%
  set_predictors(
    PLASMA_ABETA_bl, PLASMA_PTAU181_bl, PLASMA_NFL_bl,
    c(PLASMA_ABETA_bl, PLASMA_PTAU181_bl, PLASMA_NFL_bl)
  ) %>%
  set_stats(
    stat_glm(std.beta = TRUE)
  ) %>%
  fit()
#> [1] "ConvertedToAlzheimers ~ PLASMA_ABETA_bl"
#> [1] "ConvertedToAlzheimers ~ PLASMA_PTAU181_bl"
#> [1] "ConvertedToAlzheimers ~ PLASMA_NFL_bl"
#> [1] "ConvertedToAlzheimers ~ PLASMA_ABETA_bl + PLASMA_PTAU181_bl + PLASMA_NFL_bl"
#> [1] "CSF_ABETA_STATUS_bl ~ PLASMA_ABETA_bl"
#> [1] "CSF_ABETA_STATUS_bl ~ PLASMA_PTAU181_bl"
#> [1] "CSF_ABETA_STATUS_bl ~ PLASMA_NFL_bl"
#> [1] "CSF_ABETA_STATUS_bl ~ PLASMA_ABETA_bl + PLASMA_PTAU181_bl + PLASMA_NFL_bl"

fig <- model %>% aba_plot_roc()