scoreContribution: Compute the score's contribution in a network of each...

In abn: Modelling Multivariate Data with Additive Bayesian Networks

Compute the score's contribution in a network of each observation.

Description

This function computes the score's contribution of each observation to the total network score.

Usage

scoreContribution(object = NULL,
                         dag = NULL, data.df = NULL, data.dists = NULL,
                         verbose = FALSE)

Arguments

object	an object of class 'abnLearned' produced by mostProbable, searchHeuristic or searchHillClimber.
dag	a matrix or a formula statement (see details) defining the network structure, a directed acyclic graph (DAG), see details for format. Note that colnames and rownames must be set.
data.df	a data frame containing the data used for learning the network, binary variables must be declared as factors and no missing values all allowed in any variable.
data.dists	a named list giving the distribution for each node in the network, see details.
verbose	if TRUE then provides some additional output.

Details

This function computes the score contribution of each observation to the total network score. This function is available only in the mle settings. To do so one uses the glm and predict functions. This function is an attempt to perform diagnostic for an ABN analysis.

Value

A named list that contains the scores contributions: maximum likelihood, aic, bic, mdl and diagonal values of the hat matrix.

Examples

## Not run: 
## Use a subset of a built-in simulated data set
mydat <- ex1.dag.data[,c("b1","g1","p1")]

## setup distribution list for each node
mydists <- list(b1="binomial", g1="gaussian", p1="poisson")

## now build cache
mycache <- buildScoreCache(data.df = mydat, data.dists = mydists, max.parents = 2, method = "mle")

## Find the globally best DAG
mp.dag <- mostProbable(score.cache=mycache, score="bic", verbose = FALSE)

out <- scoreContribution(object = mp.dag)

## Observations contribution per network node
boxplot(out$bic)

## End(Not run)

abn documentation built on Nov. 3, 2023, 5:08 p.m.