Cluster identification with parallel coordinates

In this paper, the visualization of multi-dimensional data with the paralle l coordinate representation is studied. A particular problem of interest is the identification of lines to which clusters of points are close in an n- dimensional space. By using the duality between the Cartesian and the paral lel coordinate systems, a scan-line approach in the parallel coordinate sys tem is established and implemented to aid the identification of such lines. (C) 1999 Elsevier Science B.V. All rights reserved.