Angiogenesis Extraction Demo
The demonstration systems will automatically highlight terms and events describing angiogenesis bioprocess, as well as biological entities such as gene, gene product, tissue and cell. A Conditional Ramdom Fields (CRF) model is employed for tagging the named entities, while 3 different methods can be chosen for the extraction of angiogenesis events:
- CRF: also uses a CRF model, trained on the angiogenesis corpus.
- Keyphrase: constructs angiogenesis events using predicate argument relations generated by the ENJU syntactic parser and domain-specific key words/phrases, which were automatically extracted by comparing domain-specific and general corpora. For the moment this method is relatively slow as it calls the syntactic parser to analyse the text at run time.
- Dictionary: performs longest-match string searching using a semi-automatically compiled dictionary of phrases describing angiogenesis.
Paste a piece of text into the box below and press one of the 3 buttons. Or select:
* This is a demo version and please limit your input to 4096 characters.
Xinglong Wang |
National Centre for Text Mining |
School of Computer Science
| University of Manchester