mv_simulate: Simulate multi-view data

In andreamrau/maskmeans: Multi-view Aggregation and Splitting for K-Means Clustering

Description

Generate simulated multi-view data to illustrate the agglomerative and splitting versions of the multi-view K-means algorithm.

Usage

mv_simulate(
  beta = 4,
  n = 100,
  K = 7,
  sigma = 1.5,
  beta_V6_multiplier = 1.5,
  ...
)

Arguments

beta	Multiplicative factor controlling the spread of values around the origin in the first (beta), second (beta), and sixth (beta_V6_multiplier x beta) views. Defaults to 4.
n	Number of observations per cluster for simulations. Defaults to 100.
K	Number of clusters in the first view for simulations. As this view is made up of a single cluster at the origin surrounded by evenly spaced clusters in a circular pattern around it, this number should be odd. Defaults to 7.
sigma	Variance of noise to be added to views 2, 4, 5, and 6. Defaults to 1.5.
beta_V6_multiplier	Multiplicative parameter for the spread of values around the origin in view 6 (beta_V6_multiplier x beta). Defaults to 1.5.
...	Additional optional parameters.

Value

data	Multi-view simulated data
labels	Matrix of dimension n x v, where n is the number of observations and v the number of views, representing the true labels used to generate the data

Examples

set.seed(12345)
sim_1 <- mv_simulate(type = "D1")
sim_2 <- mv_simulate(type = "D2")
sim_3 <- mv_simulate(type = "D3")
sim_4 <- mv_simulate(type = "D4")
sim_5 <- mv_simulate(type = "D5")
sim_6a <- mv_simulate(type = "D6")
sim_6b <- mv_simulate(type = "D6", delta=7, n=200, K=5, sigma=0.5)

andreamrau/maskmeans documentation built on Nov. 13, 2021, 7:44 a.m.