Analysis of polarity information in medical text. Niu, Y., Zhu, X., Li, J., & Hirst, G. In Proceedings of the American Medical Informatics Association 2005 Annual Symposium, pages 570–574, Washington, D.C., October, 2005. abstract bibtex Knowing the polarity of clinical outcomes is important in answering questions posed by clinicians in patient treatment. We treat analysis of this information as a classification problem. Natural language processing and machine learning techniques are applied to detect four possibilities in medical text: no outcome, positive outcome, negative outcome, and neutral outcome. A supervised learning method is used to perform the classification at the sentence level. Five feature sets are constructed: UNIGRAMS, BIGRAMS, CHANGE PHRASES, NEGATIONS, and CATEGORIES. The performance of different combinations of feature sets is compared. The results show that generalization using the category information in the domain knowledge base Unified Medical Language System is effective in the task. The effect of context information is significant. Combining linguistic features and domain knowledge leads to the highest accuracy.
@InProceedings{ niu4,
author = {Yun Niu and Xiaodan Zhu and Jianhua Li and Graeme Hirst},
title = {Analysis of polarity information in medical text},
booktitle = {Proceedings of the American Medical Informatics
Association 2005 Annual Symposium},
address = {Washington, D.C.},
month = {October},
year = {2005},
pages = {570--574},
abstract = {Knowing the polarity of clinical outcomes is important in
answering questions posed by clinicians in patient
treatment. We treat analysis of this information as a
classification problem. Natural language processing and
machine learning techniques are applied to detect four
possibilities in medical text: no outcome, positive
outcome, negative outcome, and neutral outcome. A
supervised learning method is used to perform the
classification at the sentence level. Five feature sets are
constructed: UNIGRAMS, BIGRAMS, CHANGE PHRASES, NEGATIONS,
and CATEGORIES. The performance of different combinations
of feature sets is compared. The results show that
generalization using the category information in the domain
knowledge base Unified Medical Language System is effective
in the task. The effect of context information is
significant. Combining linguistic features and domain
knowledge leads to the highest accuracy.},
download = {http://ftp.cs.toronto.edu/pub/gh/Niu-etal-2005.pdf}
}
Downloads: 0
{"_id":{"_str":"521afb58aa2f288d1f000b1b"},"__v":2,"authorIDs":[],"author_short":["Niu, Y.","Zhu, X.","Li, J.","Hirst, G."],"bibbaseid":"niu-zhu-li-hirst-analysisofpolarityinformationinmedicaltext-2005","bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"firstnames":["Yun"],"propositions":[],"lastnames":["Niu"],"suffixes":[]},{"firstnames":["Xiaodan"],"propositions":[],"lastnames":["Zhu"],"suffixes":[]},{"firstnames":["Jianhua"],"propositions":[],"lastnames":["Li"],"suffixes":[]},{"firstnames":["Graeme"],"propositions":[],"lastnames":["Hirst"],"suffixes":[]}],"title":"Analysis of polarity information in medical text","booktitle":"Proceedings of the American Medical Informatics Association 2005 Annual Symposium","address":"Washington, D.C.","month":"October","year":"2005","pages":"570–574","abstract":"Knowing the polarity of clinical outcomes is important in answering questions posed by clinicians in patient treatment. We treat analysis of this information as a classification problem. Natural language processing and machine learning techniques are applied to detect four possibilities in medical text: no outcome, positive outcome, negative outcome, and neutral outcome. A supervised learning method is used to perform the classification at the sentence level. Five feature sets are constructed: UNIGRAMS, BIGRAMS, CHANGE PHRASES, NEGATIONS, and CATEGORIES. The performance of different combinations of feature sets is compared. The results show that generalization using the category information in the domain knowledge base Unified Medical Language System is effective in the task. The effect of context information is significant. Combining linguistic features and domain knowledge leads to the highest accuracy.","download":"http://ftp.cs.toronto.edu/pub/gh/Niu-etal-2005.pdf","bibtex":"@InProceedings{\t niu4,\n author\t= {Yun Niu and Xiaodan Zhu and Jianhua Li and Graeme Hirst},\n title\t\t= {Analysis of polarity information in medical text},\n booktitle\t= {Proceedings of the American Medical Informatics\n\t\t Association 2005 Annual Symposium},\n address\t= {Washington, D.C.},\n month\t\t= {October},\n year\t\t= {2005},\n pages\t\t= {570--574},\n abstract\t= {Knowing the polarity of clinical outcomes is important in\n\t\t answering questions posed by clinicians in patient\n\t\t treatment. We treat analysis of this information as a\n\t\t classification problem. Natural language processing and\n\t\t machine learning techniques are applied to detect four\n\t\t possibilities in medical text: no outcome, positive\n\t\t outcome, negative outcome, and neutral outcome. A\n\t\t supervised learning method is used to perform the\n\t\t classification at the sentence level. Five feature sets are\n\t\t constructed: UNIGRAMS, BIGRAMS, CHANGE PHRASES, NEGATIONS,\n\t\t and CATEGORIES. The performance of different combinations\n\t\t of feature sets is compared. The results show that\n\t\t generalization using the category information in the domain\n\t\t knowledge base Unified Medical Language System is effective\n\t\t in the task. The effect of context information is\n\t\t significant. Combining linguistic features and domain\n\t\t knowledge leads to the highest accuracy.},\n download\t= {http://ftp.cs.toronto.edu/pub/gh/Niu-etal-2005.pdf}\n}\n\n","author_short":["Niu, Y.","Zhu, X.","Li, J.","Hirst, G."],"key":"niu4","id":"niu4","bibbaseid":"niu-zhu-li-hirst-analysisofpolarityinformationinmedicaltext-2005","role":"author","urls":{},"metadata":{"authorlinks":{}},"html":""},"bibtype":"inproceedings","biburl":"www.cs.toronto.edu/~fritz/tmp/compling.bib","downloads":0,"keywords":[],"search_terms":["analysis","polarity","information","medical","text","niu","zhu","li","hirst"],"title":"Analysis of polarity information in medical text","title_words":["analysis","polarity","information","medical","text"],"year":2005,"dataSources":["n8jB5BJxaeSmH6mtR","6b6A9kbkw4CsEGnRX"]}