nonlinearTseries (version 0.3.1)

findAllNeighbours: neighbour search

Description

This function finds all the neighbours of all the vectors from Takens' vector array. The neighbours are found using a box assisted algorithm that creates a wrapped grid of a given number of boxes per dimension.

Usage

findAllNeighbours(takens, radius, number.boxes = NULL)

Value

A list in which the n-th position contains another list with all the neighbours of the n-th Takens' vector. If the list is empty, that means that there is no neighbour of the n-th Takens' vector in the given radius.

Arguments

takens

The matrix containing all the Takens' vectors (see buildTakens).

radius

Distance in which the algorithm will search for neighbours.

number.boxes

Integer denoting the number of boxes per dimension that will be used to construct a wrapped grid (see Schreiber). If the user does not specify a number of boxes, this function estimates a proper number.

Author

Constantino A. Garcia

References

Schreiber, T. Efficient neighbor searching in nonlinear time series analysis. Int. J. Bifurcation and Chaos, 5, p. 349, (1995).

See Also

neighbourSearch.

Examples

if (FALSE) {
# Find all the neighbours Takens' vectors build from the Henon time
# series. The size of the neighbourhood is set to 0.1.
h=henon(start = c(0.63954883, 0.04772637), do.plot = FALSE)
takens = buildTakens(h$x,embedding.dim=2,time.lag=1)
neighbours=findAllNeighbours(takens,0.1)
}