Using automated methods to handle biomedical abbreviations. Hua Xu

ISBN: 9780549858935

Published:

NOOKstudy eTextbook

158 pages


Description

Using automated methods to handle biomedical abbreviations.  by  Hua Xu

Using automated methods to handle biomedical abbreviations. by Hua Xu
| NOOKstudy eTextbook | PDF, EPUB, FB2, DjVu, AUDIO, mp3, RTF | 158 pages | ISBN: 9780549858935 | 3.33 Mb

A huge amount of information in the biomedical domain is stored as free text, such as patient records and biomedical papers. Natural language processing (NLP) technologies provide an effective way to automatically unlock information from biomedicalMoreA huge amount of information in the biomedical domain is stored as free text, such as patient records and biomedical papers.

Natural language processing (NLP) technologies provide an effective way to automatically unlock information from biomedical text. Named entity recognition (NER) is one of the biggest challenges for any biomedical NLP systems. The wide usage of biomedical abbreviations makes the NER task even more difficult. In this research, I have developed a general framework for handling biomedical abbreviations. The framework contains three components: (1) a supervised machine learning method to detect clinical abbreviations from a corpus- (2) a novel clustering method to help build corpus-specific sense inventories of clinical abbreviations- (3) a profile-based method to disambiguate ambiguous biomedical abbreviations.

Individual evaluation on each component for a corpus of clinical admission notes showed promising results. When integrated with an existing NLP system (MedLEE), the abbreviation handling system improved the performance of MedLEE by a 5.6% increase in precision and an 18.7% increase in recall. A similar profile-based disambiguation method was also applied to ambiguous gene symbols in the biomedical literature, and the evaluation showed it was very effective.

Although those methods were applied to biomedical abbreviations in this thesis, the sense detection and sense disambiguation methods are generalizable and could be used for biomedical terms in general.



Enter the sum





Related Archive Books



Related Books


Comments

Comments for "Using automated methods to handle biomedical abbreviations.":


roundbutteseed.com

©2010-2015 | DMCA | Contact us