Exploring a database through interactive visualised similarity scaling
Pieter Jan Stappers, Gert Pasman · 1999
Databases often force users to verbalize queries in terms of attribute values. In matters of taste or subjective judgement, as in picking a color, or a meal from a menu, this is difficult and distracting. In our solution, MDS-Interactive, the dialogue between user and database is mediated through a small, evolving set of sample objects, which are presented visually by a real-time multidimensional scaling (MDS) algorithm. Queries are posed by indicating positions between the samples in order to find an object having like similarities to the displayed objects. Three prototype systems have been built and evaluated.