VISUAL EXPLORATION OF COMPLEX INFORMATION SPACES BASED ON CLUSTER ANALYSIS

摘要

Searching for user-relevant information in complex information spaces, characterized by their high cardinality and dimensionality, remains a challenge for both layperson as well as expert users. Furthermore, conventional user interfaces, such as web forms, are not very helpful for a user to find information objects she needs, understand the relationships between them or get the big picture. In this work, we introduce a system, the InfoKiosk, that addresses these problems. Our approach is based on linking up techniques from the field of machine learning and information visualization rounded out by approaches for adaptivity triggered by user feedback. It allows the user to explore a complex information space by navigating in a 3D space. We show an architecture and report on first results. In particular, we give an account of our experience with game-qualified human interface devices as input devices.