Skip to contents

PFA in the tutorial is provided as scripts (R + C++ via Rcpp). This wrapper expects the PFA scripts to be bundled under inst/extdata/pfa/ in this package source. At runtime, the backend is loaded via load_backend(), which sources all .R files in that directory and compiles PFA.cpp (if present) via Rcpp::sourceCpp().

Usage

fit_pfa(
  Y_list = NULL,
  k = NULL,
  center = TRUE,
  scale = FALSE,
  nrun = 3000,
  burn = 2000,
  ...
)

Arguments

Y_list

List of study matrices.

k

Positive integer number of shared factors.

center

Logical; whether to center each variable within each study.

scale

Logical; whether to scale each variable within each study.

nrun

Number of MCMC iterations.

burn

Number of burn-in iterations.

...

Passed to PFA() in the backend script. User can set hyper-parameters and numbers of iterations here.

Value

A bifa_fit object.