full_indep: Full and independent staged event tree

In stagedtrees: Staged Event Trees

Full and independent staged event tree

Description

Build fitted staged event tree from data.

Usage

full(
  data,
  order = NULL,
  join_unobserved = TRUE,
  lambda = 0,
  name_unobserved = "UNOBSERVED"
)

## S3 method for class 'table'
full(
  data,
  order = names(dimnames(data)),
  join_unobserved = TRUE,
  lambda = 0,
  name_unobserved = "UNOBSERVED"
)

## S3 method for class 'data.frame'
full(
  data,
  order = colnames(data),
  join_unobserved = TRUE,
  lambda = 0,
  name_unobserved = "UNOBSERVED"
)

indep(
  data,
  order = NULL,
  join_unobserved = TRUE,
  lambda = 0,
  name_unobserved = "UNOBSERVED"
)

## S3 method for class 'table'
indep(
  data,
  order = names(dimnames(data)),
  join_unobserved = TRUE,
  lambda = 0,
  name_unobserved = "UNOBSERVED"
)

## S3 method for class 'data.frame'
indep(
  data,
  order = colnames(data),
  join_unobserved = TRUE,
  lambda = 0,
  name_unobserved = "UNOBSERVED"
)

Arguments

data	data to create the model, data.frame or table.
order	character vector, order of variables.
join_unobserved	logical, if situations with zero observations should be joined (default TRUE).
lambda	smoothing coefficient (default 0).
name_unobserved	name to pass to join_unobserved.

Details

Functions to create full or independent staged tree models from data. The full (or saturated) staged tree is the model where every situation is in a different stage, and thus the model has the maximum number of parameters. Conversely, the independent staged tree (indep) assigns all the situations related to the same variable to the same stage, thus it is equivalent to the independence factorization.

Examples

## full model
DD <- generate_xor_dataset(4, 100)
model_full <- full(DD, lambda = 1)

## independence model (data.frame)
DD <- generate_xor_dataset(4, 100)
model <- indep(DD, lambda = 1)
model

stagedtrees documentation built on April 29, 2022, 1:06 a.m.