tests/testthat/test_rgccak.r

In RGCCA: Regularized and Sparse Generalized Canonical Correlation Analysis for Multiblock Data

set.seed(0)
blocks <- list(
  matrix(rnorm(100 * 41), nrow = 100),
  matrix(rnorm(100 * 8), nrow = 100),
  matrix(rnorm(100 * 24), nrow = 100)
)
connection <- 1 - diag(3)
blocks <- scaling(blocks, scale = FALSE, scale_block = FALSE)
fit_rgccak <- rgccak(
  blocks, connection,
  tau = c(1, 0, 0.8), scheme = "factorial", tol = 1e-8
)

test_that("criterion is increasing", {
  n_iter <- length(fit_rgccak$crit)
  crit_old <- fit_rgccak$crit[1:(n_iter - 1)]
  crit_next <- fit_rgccak$crit[2:n_iter]
  expect_true(all(crit_next - crit_old >= 0))
})