Extracting and Visualizing Tree-Like Structures from Concept Lattices

Cassio Melo, Benedicte Le Grand, Marie-Aude Aufaure, and Anastazia Berezianos

IV’2011 conference on Information Visualization, London, July 2011.

Concept lattices built with Formal Concept Analysis are usually represented by Hasse diagrams illustrating the groupings of objects described by common attributes. Hasse diagrams display the relations of partial order between concepts in a hierarchical fashion, where each concept may have several parent concepts. Lattice visualization becomes a problem as the number of clusters grows significantly with the number of objects and attributes. Interpreting the lattice through a direct visualization of the line diagram rapidly becomes impossible and more synthetic representations are needed. In this work we propose several methods to enhance the readability of concept lattices firstly though colouring and distortion techniques, and secondly by extracting and visualizing trees derived from concept lattices structures.