nmf.rrr.DOT: DOT graph visualization for nmfae objects

In nmfkc: Non-Negative Matrix Factorization with Kernel Covariates

DOT graph visualization for nmfae objects

Description

nmfae.DOT generates a DOT language string for visualizing the structure of a three-layer NMF model. Two graph types are supported: "XCX" shows encoder factors, \Theta, and decoder factors; "YXCXY" shows the full structure from Y_2 through X_2, \Theta, X_1 to Y_1.

Edge widths are proportional to matrix element values, and edges below threshold are omitted for clarity.

Usage

nmf.rrr.DOT(
  result,
  type = c("XCX", "YXCXY"),
  threshold = 0.01,
  sig.level = 0.1,
  rankdir = "LR",
  fill = TRUE,
  weight_scale = 5,
  weight_scale_x1 = weight_scale,
  weight_scale_theta = weight_scale,
  weight_scale_x2 = weight_scale,
  Y1.label = NULL,
  X1.label = NULL,
  X2.label = NULL,
  Y2.label = NULL,
  Y1.title = "Output (Y1)",
  X1.title = "Response (X1)",
  X2.title = "Covariate (X2)",
  Y2.title = "Input (Y2)",
  hide.isolated = TRUE
)

Arguments

result	An object of class "nmfae" returned by nmfae.
type	Character. Graph type: "XCX" (default) or "YXCXY".
threshold	Numeric. Edges with values below this are omitted. Default is 0.01.
sig.level	Numeric or NULL. Significance level for filtering C edges when inference results are available. Only edges with p-value below sig.level are shown, with significance stars. Set to NULL to disable. Default is 0.1.
rankdir	Character. Graph direction for DOT layout. Default is "LR" (left to right).
fill	Logical. If TRUE, nodes are filled with color. Default is TRUE.
weight_scale	Numeric. Base scale factor for edge widths. Default is 5.
weight_scale_x1	Numeric. Scale factor for X_1 edges.
weight_scale_theta	Numeric. Scale factor for \Theta edges.
weight_scale_x2	Numeric. Scale factor for X_2 edges.
Y1.label	Character vector of output variable labels.
X1.label	Character vector of decoder basis labels.
X2.label	Character vector of encoder basis labels.
Y2.label	Character vector of input variable labels.
Y1.title	Character. Title for output node group. Default is "Output (Y1)".
X1.title	Character. Title for the response-basis node group. Default is "Response (X1)".
X2.title	Character. Title for the covariate-basis node group. Default is "Covariate (X2)".
Y2.title	Character. Title for input node group. Default is "Input (Y2)".
hide.isolated	Logical. If TRUE (default), Y1 and Y2 nodes that have no edges at or above threshold are excluded from the graph. Only applies when type = "YXCXY".

Value

A character string containing the DOT graph specification.

Lifecycle

This function is experimental. The interface may change in future versions; details are to be described in an upcoming paper.

See Also

nmfae

Examples

set.seed(1)
Y <- matrix(runif(20), nrow = 4)
res <- nmf.rrr(Y, rank1 = 2)
dot <- nmf.rrr.DOT(res)

nmfkc documentation built on July 14, 2026, 1:07 a.m.