|
| term recognition | Molecular bio. | GENIA goldstandard | Term candidates were sourced from GENIA corpus |
| term recognition | Risk management in chemical engineering | Domain experts from the Department of Chemical Engineering, Curtin University of Technology | Term candidates were extracted from the textbook "Process Systems Risk Management" |
| relation acquisition | Foods, Molecular bio. and Risk management | Initial experiments | Term candidates were manually constructed |
| web corpus construction | Risk management and Molecular bio. | Ranked based on their Odds of Domain Characterisation. The rank coefficients between the ODC based on Yahoo and Google are computed using Spearman's function. | The URLs are extracted from the respective search engines. |
| term recognition | Term recognition in Molecular Biology | Compared the performance of term recognition on four distinct corpora, namely, BootCat constructed corpus, GENIA corpus, Web corpus constructed using our approach, and simple technique of guided querying of search engines | Term candidates were sourced from GENIA corpus |
| webpage cleaning | Webpages | Evaluated using Cleaneval's development set. The gold standard is manually processed by human evaluators. Due to difficulties in using the evaluating software by Cleaneval, we employed the Text::Compare Perl module which is based on vector comparison with similarity values between 0 and 1. | From Cleaneval |
| note: as with everything else on this site, this page is best viewed using IE6+ |
