Abstract
This paper describes a simple, unsupervised bootstrapping procedure that identifies morphological description segments from heterogeneous biodiversity document collections. While the procedure is used to preprocess biodiversity literature for semantic annotation of morphological descriptions in our project, it also can be used to crawl the Web for morphological descriptions for a biodiversity niche search engine.
Original language | English (US) |
---|---|
Journal | Proceedings of the ASIST Annual Meeting |
Volume | 47 |
DOIs | |
State | Published - Nov 1 2010 |
Keywords
- Biodiversity document collections
- Morphological description
- Segment information retrieval
- Semantic annotation
- Unsupervised machine learning
ASJC Scopus subject areas
- Information Systems
- Library and Information Sciences