Computational Models of Information Scent-Following in a Very Large Browsable Text Collection

An ecological-cognitive framework of analysis and a model-tracing architecture are presented and used in the analysis of data recorded from users browsing a large document collection. The users interacted with the Scatter/Gather browser, which clusters documents into groups of similar content and presents users with summaries of cluster content. Predictions made by a computational model of navigation and information foraging are matched against the observed activity. Keywords Information foraging, information scent, cognitive modeling, model-tracing, ACT-R, Scatter/Gather

• Type: Conference
• Year: 1997
• Where: CHI
• Authors: Peter Pirolli
• Keywords: CHI 1997, Ecological-cognitive Framework, Human Computer Interaction, Information Foraging, Large Document Collection