UCSY's Research Repository

Bootstrapping Clinical Concept Extraction with Self-Training

Show simple item record

dc.contributor.author Khin, Nyein Pyae Pyae
dc.contributor.author Lynn, Khin Thidar
dc.date.accessioned 2019-07-03T07:49:11Z
dc.date.available 2019-07-03T07:49:11Z
dc.date.issued 2018-02-22
dc.identifier.uri http://onlineresource.ucsy.edu.mm/handle/123456789/297
dc.description De-identified clinical records used in this research were provided by the i2b2 National Center for Biomedical Computing and were originally prepared for the Shared Tasks for Challenges in NLP for Clinical Data. en_US
dc.description.abstract In the clinical domain, annotated clinical records are not only expensive but also often unavailable for research due to patient privacy and confidentiality requirements. The challenge is how to train effective clinical concept extraction system especially with small amount of training data. To address the limited supervision problem of insufficient labeled training examples, self-training style semi-supervised bootstrapping approach to concept extraction system is proposed. In self-training a classifier is trained from an initially small amount of human annotated data, and then used to label unlabeled data. The machine-labeled data is then added to the original data set, and the classifier is retrained iteratively. For labeling clinical concepts, Conditional Random Fields (CRF) is chosen due to its promising performance in many sequence labeling tasks. en_US
dc.language.iso en en_US
dc.publisher Sixteenth International Conferences on Computer Applications(ICCA 2018) en_US
dc.subject self-training en_US
dc.subject CRFs en_US
dc.subject semi-supervised learning en_US
dc.subject clinical concept extraction en_US
dc.title Bootstrapping Clinical Concept Extraction with Self-Training en_US
dc.type Article en_US

Files in this item

This item appears in the following Collection(s)

Show simple item record

Search Repository


My Account