ConceptScope: Organizing and Visualizing Knowledge in Documents based on Domain Ontology
share
Current text visualization techniques typically provide overviews of document content and structure using intrinsic properties such as term frequencies, term co-occurrences, and sentence structures. However, these visualization techniques do not provide conceptual overviews that consider domain-relevant knowledge that is needed when examining documents such as research articles, technical panels, or news reports. To address this shortcoming, we present ConceptScope, a text visualization technique that aids visual analysis of documents by referencing a domain ontology to represent the conceptual relationships in a document. ConceptScope uses a Bubble TreeMap visualization linked to multiple coordinated views of document structure and concept hierarchy. The visualization also includes overviews of each concept and corresponding information from the document. ConceptScope provides two visualization modes: an exploration mode that allows an overview+detail examination of a given document, and a comparison mode that allows the user to visually compare conceptual structure between multiple documents. We demonstrate ConceptScope by visualizing research articles and transcripts of technical presentations in the computer science domain, and compare it with Docuburst, the popular knowledge visualization tool for text.
READ FULL TEXT