Latent Class Discriminant Analysis

Providing a method for Local Discrimination via Latent Class Models. The approach is described in < https://www.r-project.org/conferences/useR-2009/abstracts/pdf/Bucker.pdf>.

lcda

Latent class discriminant analysis for categorical data, including local and common-components variants.

Installation

CRAN:

install.packages("lcda")

Development version:

remotes::install_github("mchlbckr/lcda")

Usage

library(lcda)

# See ?lcda, ?cclcda, and ?cclcda2 for examples

Overview

Key functions:

• lcda(): fits separate latent class models per class.
• cclcda(): fits a common-components latent class model with class-specific mixing proportions.
• cclcda2(): fits a common-components model with class-conditional mixing proportions.

Data requirements:

• Manifest variables must be integer-coded and start at 1.
• Grouping labels must be integer-coded and start at 1.

Documentation

The package includes a vignette with a worked example:

vignette("lcda")

Reference

Bücker, M., Szepannek, G., Weihs, C. (2010). Local Classification of Discrete Variables by Latent Class Models. In: Locarek-Junge, H., Weihs, C. (eds) Classification as a Tool for Research. Studies in Classification, Data Analysis, and Knowledge Organization. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-10745-0_13

Bücker, M. (2008). Lokale Diskrimination diskreter Daten. Diplomarbeit, Fakultaet Statistik, TU Dortmund.