TY - GEN
T1 - Natural Language Processing Methods for Detection of Influenza-Like Illness from Chief Complaints
AU - Hsu, Jia Hao
AU - Weng, Ting Chia
AU - Wu, Chung Hsien
AU - Ho, Tzong Shiann
N1 - Publisher Copyright:
© 2020 APSIPA.
PY - 2020/12/7
Y1 - 2020/12/7
N2 - There are several existing studies on the application of medical chief complaints in disease classification. However, the lack of a standard vocabulary and high-quality interpretation of chief complaints hinder effective classification. This study uses a variety of methods to analyze chief complaints of preschool children to detect influenza-like illness. It is expected that a fast and effective tool can be designed to assist physicians in making diagnosis, and when facing a major outbreak, it can be quickly judged to control the outbreak as soon as possible. We use several natural language processing (NLP) technologies including deep learning methods, such as the currently popular BERT model, to classify Chinese chief complaints at emergency department to detect influenza-like illness. For model evaluation, the data in 2018 were used. The method based on BERT achieved the best accuracy of 72.87% for detection of influenza-like illness.
AB - There are several existing studies on the application of medical chief complaints in disease classification. However, the lack of a standard vocabulary and high-quality interpretation of chief complaints hinder effective classification. This study uses a variety of methods to analyze chief complaints of preschool children to detect influenza-like illness. It is expected that a fast and effective tool can be designed to assist physicians in making diagnosis, and when facing a major outbreak, it can be quickly judged to control the outbreak as soon as possible. We use several natural language processing (NLP) technologies including deep learning methods, such as the currently popular BERT model, to classify Chinese chief complaints at emergency department to detect influenza-like illness. For model evaluation, the data in 2018 were used. The method based on BERT achieved the best accuracy of 72.87% for detection of influenza-like illness.
UR - http://www.scopus.com/inward/record.url?scp=85100946816&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85100946816&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85100946816
T3 - 2020 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2020 - Proceedings
SP - 1626
EP - 1630
BT - 2020 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2020 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2020 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2020
Y2 - 7 December 2020 through 10 December 2020
ER -