xhaz (version 2.1.0)

plot.predxhaz_list: Plot predictions from several xhaz models (two-panel display)

Description

This method plots a predxhaz_list object, i.e. multiple prediction objects (typically one per model and per subgroup), using two panels (left/right) for "survival" and/or "hazard".

Usage

# S3 method for predxhaz_list
plot(
  x,
  left = c("hazard", "survival"),
  right = c("survival", "hazard"),
  col = NULL,
  lty_by_group = NULL,
  legend_model_pos = c("top", "none"),
  legend_group_pos = "topright",
  ylim_left = NULL,
  ylim_right = NULL,
  grid = TRUE,
  step = NULL,
  ...
)

Value

Invisibly returns x.

Arguments

x

A predxhaz_list.

left, right

Which metrics to display: "hazard" or "survival".

col

Named colors for models (names = model names). Auto if NULL.

lty_by_group

Line types for subgroups (recycled). Auto if NULL.

legend_model_pos

"top" (outside, shared) or "none".

legend_group_pos

Inside-panel legend position (e.g., "topright") or "none".

ylim_left, ylim_right

Numeric y-limits (auto if NULL).

grid

Logical; add graphics::grid() in panels.

step

NULL = auto (step for hazard when baseline == "constant"), TRUE/FALSE to force.

...

Passed to the initial graphics::plot() in each panel (e.g., xlab, ylab, xlim).

Examples

if (FALSE) {
## --- Single-model predictions (predxhaz) ---
data("simuData", package = "xhaz")
library(survival)
fit1 <- xhaz(Surv(time_year, status) ~ agec + race,
             data = simuData,
             ratetable = survexp.us,
             interval = c(0, NA, NA, NA, NA, NA, max(simuData$time_year)),
             rmap = list(age = "age", sex = "sex", year = "date"),
             baseline = "constant", pophaz = "classic")

fit2 <- xhaz(Surv(time_year, status) ~ agec + race,
             data = simuData,
             ratetable = survexp.us,
             interval = c(0, NA, NA, max(simuData$time_year)),
             rmap = list(age = "age", sex = "sex", year = "date"),
             baseline = "bsplines", pophaz = "classic")

tgrid <- seq(0, 4, 0.1)
d1 <- simuData[1,]
px1 <- predict(fit1, new.data = d1, times.pts = tgrid, baseline = FALSE)
px2 <- predict(fit2, new.data = d1, times.pts = tgrid, baseline = FALSE)

pl <- pred_list(constant = px1, bsplines = px2)

# Two-panel plot (left/right)
plot(pl,
     left = "survival", right = "hazard",
     xlab = "Time (years)",
     ylim_left = c(0, 1))
}