GSP_NN_DISTANZ - Compute the nearest neighboor distances

Usage

: [indx, indy, dist] = gsp_nn_distanz( X1 );
[indx, indy, dist] = gsp_nn_distanz( X1, X2 );
[indx, indy, dist] = gsp_nn_distanz( X1, X2, param );
[indx, indy, dist, Xo1, Xo2, epsilon] = gsp_nn_distanz(...)

Input parameters

`X1`	Input points 1
`X2`	Input points 2
`param`	Structure of optional parameters

Output parameters

`indx`	Indices over x
`indy`	Indices over y
`dist`	Distances
`Xo1`	Points 1 after rescaling
`Xo2`	Points 2 after rescaling
`epsilon`	Radius of the ball (if the ball is used!)
`NN`	Indices of closest neighbours of each node
`Dist`	Sorted distances for each node

Description

This function computes the nearest neighboors of Xin.

Additional parameters

param.type : ['knn', 'radius'] - the type of graph (default 'knn')
param.use_flann : [0, 1] - use the FLANN library (default 0)
param.use_full : [0, 1] - Compute the full distance matrix and then sparsify it (default 0)
param.flan_checks: int - Number of checks for FLANN (default 256) the higher the more precise, but the slower. Please consider the following values: a) 32 not precise and fast, b) precise enought and still fast, c) 4096 precise and may be slow.
param.nb_cores : int - Number of cores for FLANN (default 1)
param.center : [0, 1] - center the data
param.rescale : [0, 1] - rescale the data (in a 1-ball)
param.sigma : float - the variance of the distance kernel
param.k : int - number of neighbors for knn
param.epsilon : float - the radius for the range search
param.use_l1 : [0, 1] - use the l1 distance