Skip to contents

Convert ONT signal to Gramian Angular Field

Source: R/ninetails_core_functions.R
create_gaf.Rd

Represents time series data (ONT squiggle) in a polar coordinate system instead of the typical Cartesian coordinates. This is a pure-R equivalent of the Python pyts.image implementation.

Usage

create_gaf(tail_chunk, method = "s")

Arguments

tail_chunk

Numeric vector. A 100-element signal chunk representing a fragment of the poly(A) tail within the analyzed dataset.

method

Character string specifying the type of Gramian Angular Field. "s" (default) produces a summation field (GASF), "d" produces a difference field (GADF).

Value

An array of dimensions (100, 100, 1) with values representing the Gramian Angular Field transformation of the input signal chunk. Values are in the range [0, 1].

Details

Two methods of transformation are available: Gramian Angular Summation Field (GASF) and Gramian Angular Difference Field (GADF).

The transformation proceeds as follows:

  1. Rescale signal values to the interval [-1, 1]

  2. Compute the inverse trigonometric function (acos for GASF, asin for GADF) of each value

  3. Form a pairwise matrix by summing (GASF) or differencing (GADF) the angle vectors

  4. Apply the corresponding trigonometric function (cos or sin) to the matrix

  5. Reshape to a 100x100x1 array

  6. Rescale result to the interval [0, 1]

See also

combine_gafs for combining GASF and GADF into a single array, create_gaf_list for batch GAF computation

Examples

if (FALSE) { # \dontrun{

ninetails::create_gaf(
  tail_chunk = tail_chunk,
  method = "s"
)

} # }