Fig. 2From: A novel algorithm for fast and scalable subspace clustering of high-dimensional dataProjections of dense points. Basic idea behind SUBSCALE. The projections of the points {P 7,P 8,P 9,P 10} on d 2-axis form a 1-D cluster (Cluster 3), but no 1-D cluster in dimension d 1 having the same points as Cluster 3, thus, absence of corresponding 2-dimensional cluster containing these points in the subspace {d 1,d 2}Back to article page