GQTSOM: Train a Growing Quadtree Self-Organizing Map

In EmbedSOM: Fast Embedding Guided by Self-Organizing Map

Train a Growing Quadtree Self-Organizing Map

Description

Train a Growing Quadtree Self-Organizing Map

Usage

GQTSOM(
  data,
  init.dim = c(3, 3),
  target_codes = 100,
  rlen = 10,
  radius = c(sqrt(sum(init.dim^2)), 0.5),
  epochRadii = seq(radius[1], radius[2], length.out = rlen),
  coords = NULL,
  codes = NULL,
  coordsFn = NULL,
  importance = NULL,
  distf = 2,
  nhbr.distf = 2,
  noMapping = F,
  parallel = F,
  threads = if (parallel) 0 else 1
)

Arguments

data	Input data matrix
init.dim	Initial size of the SOM, default c(3,3)
target_codes	Make the SOM grow linearly to at most this amount of nodes (default 100)
rlen	Number of training iterations
radius	Start and end training radius, as in SOM()
epochRadii	Precise radii for each epoch (must be of length rlen)
coords	Quadtree coordinates of the initial SOM nodes.
codes	Initial codebook
coordsFn	Function to generate/transform grid coordinates (e.g. tSNECoords()). If NULL (default), the grid is the grid is the 2D coordinates of GQTSOM map.
importance	Weights of input data dimensions
distf	Distance measure to use in input data space (1=manhattan, 2=euclidean, 3=chebyshev, 4=cosine)
nhbr.distf	Distance measure to use in output space (as in distf)
noMapping	If TRUE, do not compute the assignment of input data to SOM nodes
parallel	Parallelize the training by setting appropriate threads. Defaults to FALSE.
threads	Number of threads to use for training. Defaults to 0 (chooses maximum available hardware threads) if parallel=TRUE or 1 (single thread) if parallel=FALSE.

EmbedSOM documentation built on July 5, 2022, 5:14 p.m.