doubletree: doubletree

In zhenkewu/doubletree: Nested Latent Class Models for Domain-Adaptive and Semisupervised Learning of Cause-of-Deaths using Verbal Autopsy

doubletree

Description

doubletree is designed for tree-informed Bayesian domain adaptation based on nested latent class models. The special case that we address is when all the observations in a particular leaf in tree2 do not have observed leaf node label in tree1. More generally, the package can be extended to deal with

Scenario a

No missing leaf labels in tree1 or tree2, for all observations; So this is reduced to a nested latent class model estimation problem with parameter shrinkage according to the two trees.

Scenario b

All missing tree1 leaf label occurred for in a single leaf node in tree2 (say v2):

• Scenario b1: No observation with observed tree1 label in leaf v2;

• Scenario b2: More than 1 observations have observed tree1 label in leaf v2.

Scenario c

Missing tree1 leaf labels occurred for observations in 2 or more leaf nodes in tree2, say the leaf node set S:

• Sub-scenarios: 0,1,2,... leaf/leaves in S have partially observed leaf label in tree1

main double wrapper function

nlcm_doubletree()

See Also

• https://github.com/zhenkewu/doubletree for the source code and system/software requirements to use doubletree for your data.

zhenkewu/doubletree documentation built on Oct. 21, 2023, 7:04 a.m.