Medline Cognition : Pubmed searches using Semantic Natural Language Processing
Sep 23rd, 2008 by harijay
I came across a submission to Nature Precedings this morning detailing the creation of a “semantic – natural language processing ” based search for the Pubmed database by Cognition Technologies in collaboration with a team at the Department of Biochemistry at UT Southwestern Medical centre in Dallas.
A few searches later, I can definitely say the technology that drives “Medline Cognition” works and works extremely well.
The Cognition paper details the creation of its “semantic map” involving many manual steps and the improvements to existing curations, indexes and ontologies. The final result is its Medline concept index that drives its semantic search.
The paper has a brief comparison of its “Medline Cognition” searches compared to PUBMED searches. I too tried a few searches and can definitely say that unlike the other semantic web offerings that have been hyped recently, the patented-platform being built ( called CSIR) works very well and can only get better as suggested by the planned improvements detailed in their Nature Precedings paper.
[...] just found this blog post today on CODE-itch written by a “Biochemist and protein X_ray crystallographer” who [...]
Thanks for writing about Cognition! We’re glad you found our Semantic MEDLINE a helpful tool.
http://blog.cognition.com/?p=59
[...] tip to Code-Itch. Yes, I’ve had that post bookmarked since September.) [...]