R/bnmonitor.R

In bnmonitor: An Implementation of Sensitivity Analysis in Bayesian Networks

#' bnmonitor: A package for sensitivity analysis and robustness in Bayesian networks
#'
#' Sensitivity and robustness analysis for Bayesian networks.
#'
#' @details
#'
#' bnmonitor provides functions to perform sensitivity analysis for both discrete Bayesian networks (DBNs) and Gaussian Bayesian networks (GBNs). The following types of sensitivity investigations are available:
#' \itemize{
#' \item \strong{Parametric sensitivity analysis}: Investigate the effect of changes in some of the parameter values in a Bayesian network and quantify the difference between the original and perturbed Bayesian networks using dissimilarity measures (both for DBNs and GBNs).
#' \item \strong{Robustness to data}: Verify how well a Bayesian network fits a specific dataset that was used either for learning or for testing (only for DBNs).
#' \item \strong{Node influence}: Quantify how much the nodes of a Bayesian network influence an output node of interest (only for DBNs).
#' \item \strong{Edge strength}: Assess the strength of the edges of a Bayesian network (only for DBNs).
#' \item \strong{Other investigations}: Including the diameter of the conditional probability tables, measures of asymmetric independence, and level amalgamation.
#' }
#'

#' @section DBNs - Robustness to data:
#' The available functions for robustness are:
#' \itemize{
#' \item \emph{Node monitors} (\code{\link{node_monitor}}): contribution of each vertex to the overall log-likelihood of the model.
#' \item \emph{Observation's influence} (\code{\link{influential_obs}}): difference in the log-likelihood of a model learnt with the full dataset and with all but one observation.
#' \item \emph{Final node monitors} (\code{\link{node_monitor}}): marginal and conditional node monitors to assess the fit of a vertex distribution to the full dataset.
#' \item \emph{Sequential node monitors} (\code{\link{seq_node_monitor}}): marginal and conditional node monitors for a specific vertex only using sequentially subsets of the dataset.
#' \item \emph{Sequential parent-child monitor} (\code{\link{seq_pa_ch_monitor}}): parent-child node monitor for a specific vertex and a specific configuration of its parents using sequentially subsets of the dataset.
#' }
#'
#' @section DBNs - Co-variation schemes:
#' The available co-variation schemes are:
#'  \itemize{
#'   \item \emph{Uniform co-variation scheme} (\code{\link{uniform_covar}}): distributes the probability mass to be co-varied uniformly among the co-varying parameters.
#'   \item \emph{Proportional co-variation scheme} (\code{\link{proportional_covar}}): distributes the probability mass to be co-varied in the same proportion as in the original Bayesian network.
#'   \item \emph{Order-preserving co-variation scheme} (\code{\link{orderp_covar}}):distributes the to be co-varied probability mass among the co-varying parameters so that the original order of parameters is preserved.
#' }
#'
#' @section DBNs - Dissimilarity measures:
#' The dissimilarity measures quantify the difference between a Bayesian network and its update after parameter variation.\cr
#' The available dissimilarity measures are:
#' \itemize{
#'   \item \emph{Chan-Darwiche distance} (\code{\link{CD}})
#'   \item \emph{Kullback-Leibler divergence} (\code{\link{KL}})
#' }
#'
#'@section DBNs - Sensitivity functions:
#'The available functions for sensitivity analysis are:
#'  \itemize{
#'   \item \emph{Sensitivity function} (\code{\link{sensitivity}}): returns a certain probability of interest given a parameter change. Evidence can be considered.
#'   \item \emph{Sensitivity query} (\code{\link{sensquery}}): returns the parameter changes needed to get a certain probability of interest. Evidence can be considered.
#' }
#'
#'@section DBNs - Node influence:
#'The available functions for node influence are:
#'\itemize{
#'  \item \emph{Mutual information} (\code{\link{mutual_info}}): returns the mutual information between a node and all other nodes of a DBN.
#'  \item \emph{Distance-weighted influence} (\code{\link{dwi}}): returns the distance-weighted influence between a node and all other nodes of a DBN.
#'  \item \emph{Edge-weighted influence} (\code{\link{ewi}}): returns the edge-weighted influence between a node and all other nodes of a DBN.
#'}
#'
#'@section DBNs - Edge strength:
#'The available functions for edge strength are:
#'\itemize{
#'  \item \emph{Measure of edge strength} (\code{\link{edge_strength}}): returns the edge strength measure for all edges of a DBN.
#'}
#'
#'@section DBNs - Other sensitivity measures:
#'Other sensitivity measures available are:
#'\itemize{
#' \item \emph{Diameter} (\code{\link{diameter}}): returns the diameter of the conditional probability tables of all non-root nodes of a DBN.
#' \item \emph{Level amalgamation} (\code{\link{amalgamation}}): returns the diameter of all children conditional probability tables of a node in DBN when every pair of levels are merged.
#' \item \emph{Measures of asymmetric independence} (\code{\link{asy_measure}}): returns the indexes of context-specific and partial conditional independence for all variables of a DBN.
#'}
#'
#'@section GBNs - Model-Preserving matrices:
#' The available functions to construct model-preserving co-variation matrices are:
#' \itemize{
#'  \item \emph{Total co-variation matrix} (\code{\link{total_covar_matrix}}).
#'  \item \emph{Partial co-variation matrix} (\code{\link{partial_covar_matrix}}).
#'  \item \emph{Row-based co-variation matrix} (\code{\link{row_covar_matrix}}).
#'  \item \emph{Column-based co-variation matrix} (\code{\link{col_covar_matrix}}).
#' }
#'
#'@section GBNs - Mean and Covariance variations:
#' The available functions to perturb the distribution of a GBN are:
#' \itemize{
#'  \item \emph{Mean variations} (\code{\link{mean_var}}).
#'  \item \emph{Standard covariance variations} (\code{\link{covariance_var}}).
#'  \item \emph{Model-preserving covariance variations} (\code{\link{model_pres_cov}}).
#' }
#'
#'@section GBNs - Dissimilarity measures:
#' The available dissimilarity measures are:
#' \itemize{
#' \item Frobenius norm (\code{\link{Fro}}).
#' \item Jeffrey's distance (\code{\link{Jeffreys}}).
#' \item Kullback-Leibler divergence (\code{\link{KL}}).
#' \item Upper bound to the KL divergence (\code{\link{KL_bounds}}).
#' }
#'
#' @name bnmonitor
NULL