Abstract
Objective Literature-based discovery (LBD) aims to identify “hidden knowledge” in the medical literature by: (1) analyzing documents to identify pairs of explicitly related concepts (terms), then (2) hypothesizing novel relations between pairs of unrelated concepts that are implicitly related via a shared concept to which both are explicitly related. Many LBD approaches use simple techniques to identify semantically weak relations between concepts, for example, document co-occurrence. These generate huge numbers of hypotheses, difficult for humans to assess. More complex techniques rely on linguistic analysis, for example, shallow parsing, to identify semantically stronger relations. Such approaches generate fewer hypotheses, but may miss hidden knowledge. The authors investigate this trade-off in detail, comparing techniques for identifying related concepts to discover which are most suitable for LBD. Materials and methods A generic LBD system that can utilize a range of relation types was developed. Experiments were carried out comparing a number of techniques for identifying relations. Two approaches were used for evaluation: replication of existing discoveries and the “time slicing” approach.1 Results Previous LBD discoveries could be replicated using relations based either on document co-occurrence or linguistic analysis. Using relations based on linguistic analysis generated many fewer hypotheses, but a significantly greater proportion of them were candidates for hidden knowledge. Discussion and Conclusion The use of linguistic analysis-based relations improves accuracy of LBD without overly damaging coverage. LBD systems often generate huge numbers of hypotheses, which are infeasible to manually review. Improving their accuracy has the potential to make these systems significantly more usable.
Highlights
The number of academic papers being published is so large that researchers are unable to read everything potentially relevant to their research and normally focus only on publications that are directly relevant to their particular specialisation
The time segments from which these were derived are included whenever they could be found in the original paper and used for our experiments: 1. A connection between Raynaud disease and fish oil was found using Medline articles from three periods: 1983–1985,21 1980–1985,13 and 1960–1985.2 We present results from the 1960 to 1985 period
While the co-occurrence approaches clearly return a larger proportion of the gold standard, this is at the expense of generating a much larger volume of hidden knowledge over all
Summary
The number of academic papers being published is so large that researchers are unable to read everything potentially relevant to their research and normally focus only on publications that are directly relevant to their particular specialisation. A number of possible applications exist, such as: identification of treatments for diseases, drug re-purposing, disease candidate gene discovery, or drug side effect prediction.[3] For example, Swanson[4] found a connection between Raynaud’s disease and fish oil due to connecting a publication describing the effect of Raynaud’s phenomenon on blood viscosity with a separate publication containing fish oil’s effect on the same This approach to LBD, through an overlap of relationships between terms across multiple publications, is known as the A-B-C model. The patterns may be either manually created[5] or inferred from data.[6] Discovery patterns have proved useful for the discovery of novel drug applications, an application that focuses on a restricted set of concepts and clearly defined relations between them It is not clear if this technique can be applied to more open ended literature based discovery problems. Term reduction can take the form of removing frequent terms,[14] restricting target terms by the Unified Medical Language System metathesaurus (UMLS) semantic type,[15,16,17] or using association rules.[17,18] Medical Subject Heading terms have been used as underlying concepts.[19]
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: Journal of the American Medical Informatics Association : JAMIA
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
