


HIT_FIND_CLUSTERS Interface with clustering algorithms. ------------------------------------------------------------------------- DESCRIPTION ------------------------------------------------------------------------- [new_centers,new_Inl,new_class,err,clustvalid]= ... hit_find_clusters(FV,s,dim_data) ------------------------------------------------------------------------- INPUT ------------------------------------------------------------------------- FV(i,:): i-th xi-point to be clustered. s: number of clusters to be found (used just by algorithms that do not perform the estimation of the cluster number). dim_data: dimension of the xi-points. ------------------------------------------------------------------------- OUTPUT ------------------------------------------------------------------------- new_centers(i,:): center of the i-th cluster (it is a ROW vector). new_Inl: indexes of datapoints that are inliers after clustering (i.e. not discarded by the clustering algorithm). new_class(i): classification of the i-th inlier. new_class(i)=j means that the i-th inlier elongs to the j-th cluster. err: value of the clustering cost functional at the optimal solution. clustvalid: structure containing information about the clustering.results. It is equal to the structure 'spec' produced by specific clustering algorithms such as hit_kmeans or hit_single_linkage.
