Clustering is a widely used technique for data mining, indexing, and classification. We have continued our work on high-dimensional clustering, where we focus on computing projective clusters, in which points that are closely correlated in some subspace are grouped together. Instead of projecting the entire dataset on a single subspace, these methods project each cluster on its associated subspace, which is generally different from the subspace associated with another cluster. We developed a general framework called core-sets, which leads to efficient approximation algorithms for numerous problems. Using this technique we developed an efficient algorithm for certain projective clustering problems.