Other Functions

Supplementary functions for various tasks related to EntropyHub and signal processing.

---

ExampleData(SigName)

Imports sample data time series and matrices.

Data = ExampleData(SigName) 

Returns sample datasets (time series or matrices) with specific properties that are commonly used as benchmarks for assessing the performance of various entropy methods. The datasets returned by ExampleData() are used in the examples provided in documentation on www.EntropyHub.xyz and elsewhere. *Note* ExampleData() requires an internet connection to download and import the required datasets!

Data is the sample dataset imported corresponding to the string input SigName which can be one of the following:

SigName:

• One of the following strings:

uniform:

• uniformly distributed random number sequence in range [0 1], N = 5000

randintegers:

• randomly distributed integer sequence in range [1 8], N = 4096

gaussian:

• normally distributed number sequence [mean: 0, SD: 1], N = 5000

henon:

• X and Y components of the Henon attractor [alpha: 1.4, beta: .3, Xo = 0, Yo = 0], N = 4500

lorenz:

• X, Y, and Z components of the Lorenz attractor [sigma: 10, beta: 8/3, rho: 28, Xo = 10, Yo = 20, Zo = 10], N = 5917

chirp:

• chirp signal (f0 = .01, t1 = 4000, f1 = .025), N = 5000

uniform2:

• two uniformly distributed random number sequences in range [0,1], N = 4096

gaussian2:

• two normally distributed number sequences [mean: 0, SD: 1], N = 3000

randintegers2:

• two uniformly distributed pseudorandom integer sequences in range [1 8], N = 3000

uniform_Mat:

• matrix of uniformly distributed random numbers in range [0 1], N = 50 x 50

gaussian_Mat:

• matrix of normally distributed numbers [mean: 0, SD: 1], N = 60 x 120

randintegers_Mat:

• matrix of randomly distributed integers in range [1 8], N = 88 x 88

mandelbrot_Mat:

• matrix representing a Mandelbrot fractal image with values in range [0 255], N = 92 x 115

entropyhub_Mat:

• matrix representing the EntropyHub logo with values in range [0 255], N = 127 x 95

For further info on these graining procedures see the EntropyHub guide.

WindowData(Data, WinLen=None, Overlap=0, Mode='exclude')

WindowData restructures a univariate/multivariate dataset into a collection of subsequence windows.

WinData, Log = WindowData(Data) 

Windows the sequence(s) given in Data into a collection of subsequnces of floor(N/5) elements with no overlap, excluding any remainder elements that do not fill the final window. If Data is a univariate sequence (vector), Windata is a tuple of 5 vectors. If Data is a set of multivariate sequences (NxM matrix), each of M columns is treated as a sequence with N elements and WinData is a tuple of 5 matrices of size [(floor*N,5), M]. The Log dictionary contains information about the windowing process, including:

DataType:

• The type of data sequence passed as Data

DataLength:

• The number of sequence elements in Data

WindowLength:

• The number of elements in each window of WinData

WindowOverlap:

• The number of overlapping elements between windows

TotalWindows:

• The number of windows extracted from Data

Mode:

• Decision to include or exclude any remaining sequence elements (< WinLen) that do not fill the window.

WinData, Log = WindowData(Data, keyword = value, ...)

Windows the sequence(s) given in Data into a collection of subsequnces using the specified keyword arguments:

WinLen:

• Number of elements in each window, a positive integer (>10)

Overlap:

• Number of overlapping elements between windows, a positive integer (< WinLen)

Mode:

• Decision to include or exclude any remaining sequence elements (< WinLen) that do not fill the window, a string - either "include" or "exclude" (default).

See also:

ExampleData

---