pcZINB: Estimate the skeleton of the graphical model using a modified...

In yliu433/scZINB: Graphical model estimation for zero-inflated count data

Description

After estimating moral graph using neighborhood selection, the false positive connections can be removed using a modified PC algorithm, which removes an edge between two variables if they are conditional independent given any subset of other variables.

Usage

pcZINB(estG, variables, alpha = 0.01, fun = cond_ZINB)

Arguments

estG	the estimated moral graph model
variables	the sample data
alpha	the level of significance used in the conditional independence test - default is 0.01
fun	the function used for the conditional independence test

Value

A list consisting of two components:

res

the removed vertex pairs and their d-separation sets

X

simulated sample data

G

the matrix adjacency matrix for the resulting graph

yliu433/scZINB documentation built on Nov. 30, 2020, 9:07 p.m.