A 3D based user interface for information retrieval systems

Abstract

The paper describes how 3D-based visualization and interaction techniques can be used in information retrieval user interfaces. It demonstrates that information retrieval systems based on state of the art retrieval models can be supplied with an intuitive interface functionality by applying the cone-tree metaphor for the visualization of content spaces. The natural ability of humans for spatial perception, orientation and spatial memories is outlined as an advantage in the process of perceiving information spaces by means of spatial metaphors. It is shown how the models, concepts and mechanisms of the retrieval system underlying its user interface can become more transparent and perceptible for the user at the interface level and how some of the cognitive costs of navigations in information spaces can be reduced. This goal is achieved by transforming document-term networks with two levels of abstraction into hierarchical and directed cone trees that use the spatial depth of 3D to achieve an easy perception of their topological structure. Finally the paper presents a 3D-based information system user interface which gives the user intuitive control over content-oriented search paths resulting in a query generation for the underlying automatic retrieval mechanisms. In order to improve the presentation of the query results as well, it is shown that 3D-based graphical metaphors can provide very intuitive ways of perceiving the relevance of the result in accordance to the query.