Abstract
Mobile clients have limited display and navigation capabilities. To browse a set of documents, an intuitive method is to navigate through concept hierarchies. To reduce semantic loading for each term that represents the concepts and the cognitive loading of users due to the limited display, similar documents are grouped together before concept hierarchies are constructed for each document group. Since the concept hierarchies only represent the salient concepts in the documents, term extraction is necessary. Our pilot experiments showed that an unconventional combination of term frequency and inverse document frequency yielded similar performance (i.e. 71%) to previous work and the use of terms in titles achieved better performance than previous work (i.e. 82%). Our preliminary results of building concept hierarchies after clustering compared to that without is encouraging (c.f. 82% and 67%). We believe that further research can enhance the performance of concept hierarchies to a level for commercial deployment for mobile clients.
Original language | English |
---|---|
Title of host publication | Proceedings of the ACM Symposium on Applied Computing |
Pages | 627-633 |
Number of pages | 7 |
Publication status | Published - 1 Jan 2002 |
Event | Applied Computing 2002: Proceeedings of the 2002 ACM Symposium on Applied Computing - Madrid, Spain Duration: 11 Mar 2002 → 14 Mar 2002 |
Conference
Conference | Applied Computing 2002: Proceeedings of the 2002 ACM Symposium on Applied Computing |
---|---|
Country/Territory | Spain |
City | Madrid |
Period | 11/03/02 → 14/03/02 |
Keywords
- Browsing navigation
- Concept hierarchy
- Information access
- Mobile agent
- Mobile computing
- Summarization
ASJC Scopus subject areas
- Software