Benchmark Data for Alignment Methods — alignment

A small, deterministic multi-domain dataset used across alignment examples and tests. It contains three domains that share two latent classes but exhibit different linear transformations and noise levels. The object stores the raw matrices/design frames and the class labels so alignment methods can generate comparable hyperdesigns on demand.

Usage

alignment_benchmark

Format

A list with the following components:

domains: List of three elements named `domain1`, `domain2`, and `domain3`. Each element contains a centered numeric matrix `x` (samples x features) and a `design` data frame with `id`, `condition`, and `domain` columns.
labels: Factor of length 80 giving the latent class for each sample.

Examples

data <- alignment_benchmark
lapply(data$domains, function(dom) dim(dom$x))
#> $domain1
#> [1] 80  4
#> 
#> $domain2
#> [1] 80  4
#> 
#> $domain3
#> [1] 80  4
#> 
table(data$labels)
#> 
#> class_A class_B 
#>      40      40