BioGSP (version 1.0.0)

runSGCC: Run SGCC weighted similarity analysis in Fourier domain

Description

Calculate energy-normalized weighted similarity between two signals using Fourier domain coefficients directly (no vertex domain reconstruction). Excludes DC component and uses energy-based weighting consistent with Parseval's theorem.

Usage

runSGCC(
  signal1,
  signal2,
  SG = NULL,
  eps = 1e-12,
  validate = TRUE,
  return_parts = TRUE,
  low_only = FALSE
)

Value

Similarity analysis results computed in Fourier domain

Arguments

signal1

Either a signal name (character) for SG object, or SGWT Forward result, or SGWT object

signal2

Either a signal name (character) for SG object, or SGWT Forward result, or SGWT object

SG

SGWT object (required if signal1/signal2 are signal names)

eps

Small numeric for numerical stability (default: 1e-12)

validate

Logical; if TRUE, check consistency (default: TRUE)

return_parts

Logical; if TRUE, return detailed components (default: TRUE)

low_only

Logical; if TRUE, compute only low-frequency similarity (default: FALSE)

Examples

# \donttest{
# Create example data and compute SGWT
data <- data.frame(x = runif(100), y = runif(100),
                  signal1 = rnorm(100), signal2 = rnorm(100))
SG <- initSGWT(data, signals = c("signal1", "signal2"))
SG <- runSpecGraph(SG, k = 15)
SG <- runSGWT(SG)

# Between two signals in same SGWT object
similarity <- runSGCC("signal1", "signal2", SG = SG)
print(similarity)

# Between two SGWT objects
data2 <- data.frame(x = runif(100), y = runif(100), signal = rnorm(100))
SG2 <- initSGWT(data2, signals = "signal")
SG2 <- runSpecGraph(SG2, k = 15)
SG2 <- runSGWT(SG2)

similarity2 <- runSGCC(SG, SG2)
print(similarity2)
# }