reweight: Re-fitting Models with New Weights

In party: A Laboratory for Recursive Partytioning

Re-fitting Models with New Weights

Description

Generic function for re-fitting a model object using the same observations but different weights.

Usage

reweight(object, weights, ...)

Arguments

object	a fitted model object.
weights	a vector of weights.
...	arguments passed to methods.

Details

The method is not unsimilar in spirit to update, but much more narrowly focused. It should return an updated fitted model derived from re-fitting the model on the same observations but using different weights.

Value

The re-weighted fitted model object.

See Also

update

Examples

  ## fit cars regression
  mf <- dpp(linearModel, dist ~ speed, data = cars)
  fm <- fit(linearModel, mf)
  fm
  
  ## re-fit, excluding the last 4 observations
  ww <- c(rep(1, 46), rep(0, 4))
  reweight(fm, ww)

party documentation built on Sept. 11, 2024, 5:18 p.m.