Skip to contents

Multi-Graph GRASP: Graph Alignment by Spectral Corresponding Functions

Source: R/all_generic.R, R/grasp_multiset.R
grasp_multiset.Rd

Performs GRASP alignment on three or more graph domains simultaneously. Extends the pairwise GRASP algorithm to handle multiple graphs through joint diagonalization and shared latent basis alignment.

Usage

grasp_multiset(data, ...)

grasp_multiset(data, ...)

# S3 method for class 'hyperdesign'
grasp_multiset(
  data,
  preproc = multivarious::center(),
  ncomp = 30,
  q_descriptors = 100,
  sigma = 0.73,
  lambda = 0.1,
  use_laplacian = TRUE,
  anchor = 1L,
  solver = c("linear", "auction"),
  assignment_alpha = 0.9,
  max_iter = 100,
  tol = 1e-06,
  ...
)

# S3 method for class 'list'
grasp_multiset(data, ...)

# Default S3 method
grasp_multiset(data, ...)

Arguments

data

A list containing three or more matrices (nodes x features)

...

Additional arguments passed to hyperdesign method

preproc

Preprocessing step applied via multivarious::init_transform

ncomp

Number of spectral components (default: 30)

q_descriptors

Number of descriptors (default: 100)

sigma

Diffusion parameter (default: 0.73)

lambda

Regularization parameter (default: 0.1)

use_laplacian

Whether to use Laplacian normalization (default: TRUE)

anchor

Index of reference graph (1-based) or "mean" (default: 1)

solver

Assignment solver: "auction" or "linear" (default: "auction")

assignment_alpha

Weight for cosine vs. Euclidean assignment cost (0-1)

max_iter

Maximum iterations for joint optimization (default: 100)

tol

Convergence tolerance (default: 1e-6)

Value

A grasp_multiset object containing alignment results for all graphs

A grasp_multiset object

See also

grasp for pairwise alignment, grasp_multiset.hyperdesign for detailed documentation

Examples

# \donttest{
set.seed(1)
X1 <- matrix(rnorm(30), 10, 3)
X2 <- matrix(rnorm(30), 10, 3)
X3 <- matrix(rnorm(30), 10, 3)
data_list <- list(X1, X2, X3)
result <- grasp_multiset(data_list, ncomp = 5)
#> Warning: Joint alignment reached maximum iterations without convergence
str(result$embeddings)
#> List of 3
#>  $ : num [1:10, 1:5] -0.167 -0.271 -0.182 0.356 -0.057 ...
#>  $ : num [1:10, 1:5] 1.8 -1.84 -1.55 -0.44 1.73 ...
#>  $ : num [1:10, 1:5] -1.3831 -1.3723 -0.1632 1.4176 0.0708 ...
# }