MODECLUSTF Fast KNN modeclustering, based on overlaping cells
[LAB,NNLAB,NDIST] = MODECLUSTF(A,C,K,NEST)
DescriptionThis is a fast version of MODECLUST, useful and essential for very large dataset (more than a million objects). It makes for every object a rough estimate of the potential set of nearest neighbors (which should be larger than max(K)). This set is larger for larger values of C, resulting in a slower, but more accurate procedure. In many practical problems it appeared that C = 6 was sufficient. The computational complexity of this algorithm (number of distances that are actually computed) is M x SQRT(M). However, for small M it is not faster than MODECLUST. Therefor, it will jump to that routine for small values of M as well as for C == 0. Reference(s)R.P.W. Duin and S. Verzakov, Fast kNN mode seeking clustering applied to active learning, arXiv:1712.07454, 2017, 123. See alsomappings, datasets, distm, proxm, clustm, dclustm, modeclust_batch, modeclust, reclustn,
