select: Select solutions from a solution path

In sparsebnUtils: Utilities for Learning Sparse Bayesian Networks

Description

Choose solutions from a solution path based on number of edges, value of regularization parameter lambda, or index.

Usage

select(x, edges, lambda, index)

Arguments

x	a sparsebnPath object.
edges	number of edges to search for.
lambda	value of regularization parameter to search for.
index	integer index to select.

Details

For edges (resp. lambda), the solution with the closest number of edges (resp. regularization parameter) is returned. If there is no match within a tolerance of 0.1 for lambda, nothing is returned. Fuzzy matching is not used for when selecting by index.

If there is more than one match (for example, by number of edges), then the first such estimate is returned. Note that select(x, index = j) is equivalent to (but slightly slower than) x[[j]].

sparsebnUtils documentation built on Jan. 27, 2021, 9:05 a.m.