Sentiment classification with medical word embeddings and sequence representation for drug reviews
Conference Publication ResearchOnline@JCUAbstract
Medical sentiments derived from health-care related documents,such as health reviews, tweets or forums, have been an indispensable resource for studying insights into patient health conditions and generating additional information for health professionals to provide more supportive treatments. However, approaches implemented in previous studies indicate inadequacy in discovering insights into review details and implicit emotional information due to domain specificities. We propose a sentiment classification framework with medical word embeddings and sequence representation for drug review datasets. Empirical results on different vector transformation methods imply the superiority of sequence incorporated medical sentiment lexicon using machine learning classifiers. Experiments on various word embeddings with convolutional neural network model further justify the effectiveness of medical sentiment word embeddings in sentiment classification for drug reviews.
Journal
N/A
Publication Name
HIS 2018: 7th International Conference on Health Information Science
Volume
11148
ISBN/ISSN
978-3-030-01077-5
Edition
N/A
Issue
N/A
Pages Count
12
Location
Cairns, QLD, Australia
Publisher
Springer
Publisher Url
N/A
Publisher Location
Cham, Switzerland
Publish Date
N/A
Url
N/A
Date
N/A
EISSN
N/A
DOI
10.1007/978-3-030-01078-2_7