Fit a distribution to truncated severity (loss) data — fit_truncated

Estimate the original distribution from truncated data.

Usage

fit_truncated_dist(
  y,
  dist = c("gamma", "lognormal"),
  left = NULL,
  right = NULL,
  start = NULL,
  print_initial = TRUE,
  n_variants = 1,
  n_shape_grid = 8,
  n_scale_grid = 8,
  trace = FALSE,
  report = TRUE
)

Arguments

y: Vector with observations of losses.
dist: Distribution for severity: "gamma" or "lognormal".
left: Numeric. Observations below this threshold are not present in the sample.
right: Numeric. Observations above this threshold are not present in the sample.
start: Optional list of starting values. If NULL, a multi-start strategy is used.
print_initial: Deprecated logical retained for backward compatibility.
n_variants: Controls how many local variations around base starts are used.
n_shape_grid: Number of grid points for gamma shape.
n_scale_grid: Number of grid points for gamma scale.
trace: Logical. If TRUE, prints periodic progress during the fitting loop.
report: Logical. If TRUE, prints a short summary at the end.

Value

An object of class c("truncated_dist", "fitdist") with additional attributes describing the attempted fits.