GreedySearchParentsWithPriors: This function applied greedy hill climbing search to find the...

In HamdaBinteAjmal/DBN4GRN: An R Package for Learning Sparse DBN Structures to Reverse Engineer Gene Regulatory Networks

Description

This function applied greedy hill climbing search to find the parents of each node in the network from the previous time slice. The score for all arcs absent in the network is calculated as :

Usage

GreedySearchParentsWithPriors(
  data,
  score_mat = NULL,
  maxP = Inf,
  gamma = 0,
  pred_pos = NULL,
  score = "Score_BIC"
)

Arguments

data	a data.frame where each column represents the variables and rows contain the temporal data
score_mat	a matrix of scores of edges to be added to the BIC score. This can be the matrix of lasso regression coefficience or the p-value of the first order conditional independence tests.
maxP	maximum number of allowed parents per node
gamma	value of the gamma constant added to the final score, to move the very small edge score values away from zero
pred_pos	position of the putative regulators in the dataset. If this value is NULL, all variables are considered as putative regulators.
score	This refers to the type of score to be calculated. The default is "Score_LOPC". Other possible values are Score_LASSO and BIC.

Value

bnlearn object

HamdaBinteAjmal/DBN4GRN documentation built on Dec. 27, 2021, 2:08 p.m.