Font Size: a A A

Biomedical Entity Annotation Using Linked Open Data

Posted on:2015-04-22Degree:MasterType:Thesis
Country:ChinaCandidate:L TianFull Text:PDF
GTID:2298330452964024Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
With the ever-growing use of textual biomedical data, domain entity annotationhas become very important in biomedicine. It is quite useful for many tasks includ-ing information extraction, classifcation, text summarization, question answering, andliterature-based knowledge discovery.Although previous work has been widely used inannotating domain entities from biomedical references, it sufers from several issues,suchasadatafexibilityproblem, languagedependencyandthedifcultyofrule-basedinference for word sense disambiguation.Ontheotherhand,theLinkedOpenData(LOD)Initiativeaimsatinterlinkingdatafromvariousopenknowledgebases. Inthelastfewyears, theamountofstructureddataavailableon the Webhas been increasing rapidly. Currently, there are billionsof triplespublicly available in295Web data sources of diferent domains written in a variety oflanguages. The numbers of entities and properties describing semantic relationshipsbetween entities within the linked data cloud have become very large. It is high-yieldfor us to get specifc area of data from the cloud.In this paper, we propose a knowledge-incentive approach for entity annotationin biomedical feld based on LOD. With the approach, we implement MeDetect, aprototype systemto solve the problems mentioned above, especiallyentity fltering anddisambiguation by using collective annotation based on Linked Open Data. Finally,we present the results of experiments that verify the efectiveness and efciency of ourapproach.
Keywords/Search Tags:Domain Entity Annotation, Linked Open Data, Bio-Informatics
PDF Full Text Request
Related items