Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Review
. 2009 Apr;42(2):390-405.
doi: 10.1016/j.jbi.2009.02.002. Epub 2009 Feb 14.

Empirical Distributional Semantics: Methods and Biomedical Applications

Affiliations
Free PMC article
Review

Empirical Distributional Semantics: Methods and Biomedical Applications

Trevor Cohen et al. J Biomed Inform. .
Free PMC article

Abstract

Over the past 15 years, a range of methods have been developed that are able to learn human-like estimates of the semantic relatedness between terms from the way in which these terms are distributed in a corpus of unannotated natural language text. These methods have also been evaluated in a number of applications in the cognitive science, computational linguistics and the information retrieval literatures. In this paper, we review the available methodologies for derivation of semantic relatedness from free text, as well as their evaluation in a variety of biomedical and other applications. Recent methodological developments, and their applicability to several existing applications are also discussed.

Figures

Figure 1
Figure 1
Visualization of the 15 nearest neighbors of “thrombosis” in the OHSUMED corpus using SVD to reduce the neighborhood to the two dimensions that best capture the variance between the original high-dimensional points. The third most significant dimension is encoded as color and font size. Note the clustering of “thrombus” and “thrombi”, as well as “venography”, “venous” and “vein”.
Figure 2
Figure 2
Visualization of the same 15 nearest neighbors of “thrombosis” using Pathfinder network scaling to preserve the most significant links in the network of terms connected by their semantic similarity. Note the preservation of the connection between “thrombi” and “thrombus”, as well as between “venous” and “vein” and “heparin” (an anticoagulant) and “anticoagulation”.

Similar articles

See all similar articles

Cited by 45 articles

See all "Cited by" articles

Publication types

LinkOut - more resources

Feedback