Trelliscopejs Display for Multiple Imputation Diagnostics

With the support of trelliscopejs, vismi provides trelliscope displays to inspect all variables at once via

• trellis_vismi()

• trellis_vismi_overimp()

• trellis_vismi_converge()

library(vismi)
library(mice)
library(mixgb)

Distributional characteristics via trellis_vismi()

trellis_vismi(data = newborn, imp_list = imp_newborn, marginal_x = "box+rug")

Overimputation diagnostics via trellis_overimp()

obj <- overimp(data = newborn, p = 0.2, m = 5, test_ratio = 0.2,
    method = "mixgb")

trellis_vismi_overimp(obj = obj, stack_y = TRUE)

Convergence diagnostics via trellis_vismi_converge()

library(mixgb)
set.seed(2026)
mixgb_obj <- mixgb(data = newborn, m = 5, maxit = 5, save.models = TRUE)

trellis_vismi_converge(obj = mixgb_obj)

library(mice)
set.seed(2026)
mice_obj <- mice(data = newborn, m = 5, maxit = 5, printFlag = FALSE)

trellis_vismi_converge(obj = mice_obj)