Search for FNN to get an optimal Embedding Dimension using by using nonlinearTseries::findAllNeighbours() in a loop.
fnn(y, emLag = 1, maxDim = 10, radius = sd(y)/10, number.boxes = NULL)A numeric vector or time series
Optimal embedding lag (delay), e.g., provided by an optimising algorithm. If NULL the lags based on the mutual information in lagMethods will be reported. If a numeric value representing a valid lag is passed, this value will be used to estimate the number of dimensions (default = NULL)
Maximum number of embedding dimensions to consider (default = 10)
Size of the neighbourhood: Every point smaller than the radius will be considered a near neighbour, see nonlinearTseries::findAllNeighbours() (default = sd(y)/10).
Integer representing number of boxes to to speed up neighbour search, if NULL an optimal number will be chosen nonlinearTseries::findAllNeighbours() (default = NULL).
FNN curve