Revised Papers from the Second Kyoto Workshop on Digital Cities II, Computational and Sociological Approaches
Theories of cognition and image categorization: What category labels reveal about basic level theory
Journal of the American Society for Information Science and Technology
| Hi-index | 0.00 |
Information visualization can improve the access and manipulation of digital data if it conforms to human expectations, and supports the particular interface, task, data, and user. Detailed usability studies are required to prove this assumption. This paper reports a usability study of the data analysis algorithm applied in the LVis - Digital Library Visualizer that spatially visualizes search results derived from user queries of digital library collections. In particular, it presents a comparison of free sorting results for image data done by 20 human subjects and the data analysis result derived via Latent Semantic Analysis from the textual image descriptions. The results of the study suggest that people, if asked to categorize with no constraints, will use features that are often derivable from textual metadata. Human clustering strategies and resulting demands for image browsing systems are discussed and future work is outlined.