An Algorithm for l-diversity Clustering of a Point-Set

Clustering a set of data means to divide the data into some disjoint groups on the basis of some properties of the entries. Given a set of n points plotted in a metric space where each of them have a color, an l-diversity clustering algorithm clusters the points into a set of disjoint clusters such that each cluster has
$l$
points and all points in a cluster have distinct colors. To produce clusters with better qualities, we have proposed an l-diversity clustering algorithm which minimizes the length of the maximum radius of all the clusters, the sum of radii of all clusters and the amount of data loss. We have compared the outputs of our algorithm with that of the previously best known l-diversity clustering algorithm for the same input data sets. For each input data set, our algorithm produced better clustering by minimizing the maximum radius, the sum of radii and the data loss.