ANF: Fuse affinity networks (i.e., matrices) through one-step or...

In ANF: Affinity Network Fusion for Complex Patient Clustering

Description

Fuse affinity networks (i.e., matrices) through one-step or two-step random walk

Usage

ANF(Wall, K = 20, weight = NULL, type = c("two-step", "one-step"),
  alpha = c(1, 1, 0, 0, 0, 0, 0, 0), verbose = FALSE)

Arguments

Wall	a list of affinity matrices of the same shape.
K	the number of k nearest neighbors for function kNN_graph
weight	a list of non-negative real numbers (which will be normalized internally so that it sums to 1) that one-to-one correspond to the affinity matrices included in 'Wall'. If not set, internally uniform weights are assigned to all affinity matrices in 'Wall'.
type	choose one of the two options: perform "one-step" random walk, or "two-step" random walk on the list of affinity matrices in 'Wall“ to generate a fused affinity matrix. Default: "two-step" random walk
alpha	a list of eight non-negative real numbers (which will be normalized internally to make it sums to 1). Only used when "two-step" (default value of 'type') random walk is used. 'alpha' is the weights for eight terms in the "two-step" random walk formula (check research paper for more explanations about the terms). Default value: (1, 1, 0, 0, 0, 0, 0, 0), i.e., only use the first two terms (since they are most effective in practice).
verbose	logical(1); if true, print some information

Value

a fused transition matrix (representing a fused network)

Examples

D1 = matrix(runif(400), nrow=20)
W1 = affinity_matrix(D1, 5)
D2 = matrix(runif(400), nrow=20)
W2 = affinity_matrix(D1, 5)
W = ANF(list(W1, W2), K=10)

ANF documentation built on Nov. 8, 2020, 7:51 p.m.