|
1000
|
Abstract/Summary
|
-
The Information Retrieval Systems in the Life Sciences focus on Medline citations as primary source of informa0tion. This is reasoned by the situation that Medline citations pretend to contain the most important parts of a publication. The knowledge environment SCAIView is extended to analysis functionality to extract easier the information content. SCAIView’s corpus is extended from Medline abstracts to full texts and patents. Full texts are taken from PubMed Central and patents are taken from the TREC Chemistry Track. SCAIView is compared to other Information Retrievals based on reference gene lists related to Parkinson Disease, Alzheimer Disease, Schizophrenia and Intracranial Aneurysm. The performance analyses show that SCAIView is the best Information Retrieval System in the Life Sciences because of its high adjustment features. These performance analyses are applied on the two new corpora: PubMed Central full texts and TREC patents. The results of the analyses on these two corpora reveal that the information content in patents and full texts is much higher than in Medline citations. Patents and full texts contain much more information than abstracts, especially because of their tables.
|
