fractal (version 2.0-4)

embedSeries: Creates a delay embedding of a single variable time series

Description

Given the time series \(X_t\), the embedding dimension \(E\), and the time delay \(\tau\), the embedding coordinates are defined as \(X_t, X_{t-\tau},\ldots,X_{t - (E-1)\tau}\).

Usage

embedSeries(x, dimension=NULL, tlag=NULL, series.name=NULL)

Arguments

x

a vector containing a uniformly-sampled real-valued time series.

dimension

the maximal embedding dimension. Default: 2.

series.name

a character string defining the name of the input time series. Default: deparseText(substitute(x)).

tlag

the time delay between coordinates. Default: the first zero crossing of the autocorrelation function of x.

Value

an object of class embedSeries.

S3 METHODS

[

data access method. Usage: x[1:3,1].

as.matrix

convert embedding into matrix object. Usage: as.matrix(x).

eda.plot

creates an extended data analysis plot of the data summarizing many of its statistical features. Usage: eda.plot(x).

plot

plots the embedding. For embeddings higher than 3, a spin plot of the data is generated. Use the buttons on the spin control panel to conrtol the display. Available options to the plot function are:

dim

The plot dimension. Must be less than or equal to the maximal embedding dimension (number of columns in the embedding matrix). Default: the maximal embedding dimension.

...

Additional plot arguments (set internally by the par function).

Usage: plot(x).

print

prints a summary of the embedding. Available options are:

...

Additional print arguments used by the standard print function.

Usage: print(x).

See Also

timeLag, FNN, corrDim, determinism.

Examples

# NOT RUN {
## embed the beamchaos series in 10 dimensions 
## using a time lag of 15. 
z <- embedSeries(beamchaos, tlag=15, dim=10)

## plot the attractor in the phase space 
# }
# NOT RUN {
plot(z)
# }
# NOT RUN {
## plot the embedding projected down to two 
## dimensions 
plot(z, dim=2)
# }