Edit an existing smoothing step in a refinement workflow

Manually adjusts a smoothing step that was previously added with add_smoothing(). This is intended for actuarial review of a smoothed tariff curve, for example to flatten an unstable segment, align the end points of an interval, or add extra control points where expert judgement should guide the curve. The adjusted smoothing is applied when refit() is called.

Usage

edit_smoothing(
  model,
  model_variable = NULL,
  step = NULL,
  from,
  to,
  from_value = NULL,
  to_value = NULL,
  control_positions = NULL,
  control_values = NULL,
  allow_extrapolation = FALSE,
  extrapolation_step = NULL
)

Arguments

model

Object of class rating_refinement, usually created with prepare_refinement(). Legacy smooth and restricted objects are still accepted for backwards compatibility.

model_variable

Character string. The model_variable of the smoothing step to edit. Required when more than one smoothing step exists and step is not supplied.

step

Optional numeric index of the smoothing step to edit.

from, to

Numeric values giving the start and end of the source-variable interval to modify.

from_value, to_value

Optional numeric values used to override the smoothed curve value at from and to.

control_positions, control_values

Optional numeric vectors of equal length. These define additional points that the edited smoothing curve should pass through.

allow_extrapolation

Logical. Whether edits may extend beyond the observed source-variable range.

extrapolation_step

Optional positive numeric scalar used to set the spacing of extra break points when extrapolation is allowed.

Value

Object of class rating_refinement.

Details

Use model_variable or step to identify the smoothing step to edit. The interval from from to to defines the part of the source variable range that should be changed. from_value and to_value can be used to force the curve values at the interval boundaries. control_positions and control_values add additional points that the edited curve should follow inside the interval.

Author

Martin Haringa

Examples

set.seed(42)
driver_age <- rep(seq(20, 59), each = 4)
exposure <- rep(1, length(driver_age))
age_band <- cut(
  driver_age,
  breaks = c(18, 30, 40, 50, 60),
  include.lowest = TRUE
)
expected_claims <- exp(
  -1.7 + 0.018 * (driver_age - 20) + 0.0006 * (driver_age - 40)^2
)
portfolio <- data.frame(
  claims = rpois(length(driver_age), exposure * expected_claims),
  exposure = exposure,
  driver_age = driver_age,
  age_band = age_band
)

model <- glm(
  claims ~ age_band + offset(log(exposure)),
  family = poisson(),
  data = portfolio
)

refined <- prepare_refinement(model, data = portfolio) |>
  add_smoothing(
    model_variable = "age_band",
    source_variable = "driver_age",
    breaks = c(18, 30, 40, 50, 60),
    degree = 2,
    weights = "exposure"
  ) |>
  edit_smoothing(
    model_variable = "age_band",
    from = 30,
    to = 50,
    from_value = 1.00,
    to_value = 1.10,
    control_positions = c(40),
    control_values = c(1.05)
  )

refined_model <- refit(refined)