mosaic_sim: Simulate a multi-source data set with planted clusters

In MosaiClusteR: An Umbrella Framework for Multi-Source and Multi-Omics Clustering

Simulate a multi-source data set with planted clusters

Description

Generates K feature-by-sample matrices that share a common set of n samples drawn from g ground-truth groups. Each source has its own informative features (shifted between groups) plus noise features, so the clusters are only partially visible in any single source - the canonical setting in which multi-source integration helps.

Usage

mosaic_sim(
  n = 60,
  g = 3,
  p = c(120, 150),
  informative = c(20, 25),
  effect = 2.5,
  K = 2,
  seed = NULL
)

Arguments

n	Number of samples (objects).
g	Number of ground-truth groups.
p	Per-source feature counts (recycled to length K).
informative	Per-source number of group-informative features.
effect	Between-group mean shift for informative features.
K	Number of sources.
seed	Optional RNG seed.

Value

A list with List (the K matrices, objects in rows, features in columns) and truth (the integer group label per object).

Examples

sim <- mosaic_sim(n = 30, g = 3, K = 2, seed = 1)
lengths(lapply(sim$List, dim))

MosaiClusteR documentation built on Oct. 25, 2026, 1:08 a.m.