TS2IMF: Decompose one time series into intrinsic mode functions.

In gmsp: Ground Motion Signal Processing

Decompose one time series into intrinsic mode functions.

Description

Orchestrates VMD/EMD/EEMD decomposition for one canonical t/s signal and returns either IMFs, a recomposed time series, or long/wide tables depending on output.

TS2IMF() is a worker. It does not detect or own record/component grouping. Use data.table grouping at the call site for TSL tables.

Usage

TS2IMF(
  .x,
  method = "vmd",
  K = 12,
  alpha = 2000,
  tau = 0,
  DC = TRUE,
  init = 0,
  tol = 1e-07,
  output = NULL,
  verbose = FALSE,
  boundary = "wave",
  stop.rule = "type5",
  noise.type = "gaussian",
  noise.amp = 5e-08,
  trials = 10,
  imf.remove = NULL,
  remove.Fo = NULL,
  keep.Fo = NULL,
  keep.Residue = TRUE
)

Arguments

.x	data.table with canonical t and s columns.
method	character. One of "vmd", "emd", "eemd" (default "vmd").
K	integer. Number of IMFs (default 12).
alpha, tau, DC, init, tol	numeric. Parameters for VMD.
output	character or NULL. One of "TSL", "TSW", "IMF" (default NULL).
verbose	logical. Engine verbosity (default FALSE).
boundary, stop.rule	character. EMD/EEMD parameters.
noise.type, noise.amp, trials	parameters for EEMD.
imf.remove	character or integer. IMF selection (optional). Character values remove explicit mode names such as "IMF1". Numeric values select IMF indices to remove: positive indices count from the first IMF, negative indices count from the last IMF, and both sets are unioned. For example, c(1L, -1L) removes IMF1 and the last IMF. Zero, non-finite, and out-of-range numeric values are ignored.
remove.Fo, keep.Fo	numeric length-2 (Hz) frequency band rules (optional).
keep.Residue	logical. If TRUE (default), include residue in the reconstruction of signal.R. If FALSE, signal.R is built only from the kept IMFs.

Value

Depending on output, returns TSL, TSW, IMF or a list with the decomposition.

Examples

t <- seq(0, 1, by = 0.01)
x <- data.table::data.table(
  t = t,
  s = sin(2 * pi * t) + 0.1 * sin(10 * pi * t)
)
imf <- TS2IMF(x, method = "vmd", K = 2, output = "IMF")
imf

gmsp documentation built on July 18, 2026, 5:07 p.m.