LAWBL: Latent (variable) Analysis with Bayesian Learning

Installation

1. Install the stable version from CRAN with:

install.packages("LAWBL")

2. Install the devtools package (if necessary), and install the development version from the Github.

# install.packages("devtools")
devtools::install_github("Jinsong-Chen/LAWBL")

What can this package do?

The long-term goal of LAWBL is to provide a analytical framework for modeling latent variables with different Bayesian learning methods.

Currently, this package includes the Partially Confirmatory Factor Analysis (PCFA), a partially confirmatory approach covering a wide range of the exploratory-confirmatory continuum in factor analytic models (Chen, Guo, Zhang, & Pan, 2020). There are two major model variants with different constraints for identification. One assumes local independence (LI) with a more exploratory tendency, which can be also called the E-step. The other allows local dependence (LD) with a more confirmatory tendency, which can be also called the C-step. Parameters are obtained by sampling from the posterior distributions with the Markov chain Monte Carlo (MCMC) techniques. Different Bayesian Lasso methods are used to regularize the loading pattern and local dependence.

Although only continuous data are supported currently, inclusion of mixed-type data is on schedule. More Bayesian learning approaches will be also included in future releases of this package.

For examples of how to use the package, see vignettes or here.

How to use this package?

• To estimate the E-step (when only a few loadings can be specified, e.g., 2 per factor), use m <- pcfa(dat=dat,Q=Q,LD=F)
• To estimate the C-step (with one specified loading per item), use m <- pcfa(dat=dat,Q=Q,LD=T)
• To summarize basic information after estimation, use summary(m)
• To summarize significant loadings in pattern/Q-matrix format, use summary(m,what=‘qlambda’)
• To summarize factorial eigenvalues, use summary(m,what=‘eigen’)
• To summarize significant LD terms, use summary(m,what=‘offpsx’)
• To plot eigenvalues’ trace, use plot_eigen(m)
• To plot eigenvalues’ density, use plot_eigen(m, what=‘density’)
• To plot eigenvalues’ adjusted PSRF, use plot_eigen(m, what=‘APSR’)

Reference

Chen, J., Guo, Z., Zhang, L., & Pan, J. (2020). A partially confirmatory approach to scale development with the Bayesian Lasso. Psychological Methods. Advance online publication. http://dx.doi.org/10.1037/met0000293