Recognition of side effects as implicit-opinion words in drug reviews

TitleRecognition of side effects as implicit-opinion words in drug reviews
Publication TypeJournal Article
Year of Publication2016
AuthorsMonireh Ebrahimi, Amir Hossein Yazdavar, Naomie Salim, Safaa Eltyeb
Volume40
Issue7
Start Page1018
Pagination1018-1032
Date Published2016/11/14
PublisherOnline Information Review
Type of ArticleJournal
KeywordsDrug review, Drug side effect, Medical-opinion mining, Regular expression, Rule based, Supervised approach
Abstract

Purpose – Many opinion-mining systems and tools have been developed to provide users with the
attitudes of people toward entities and their attributes or the overall polarities of documents.
In addition, side effects are one of the critical measures used to evaluate a patient’s opinion for a
particular drug. However, side effect recognition is a challenging task, since side effects coincide with
disease symptoms lexically and syntactically. The purpose of this paper is to extract drug side effects
from drug reviews as an integral implicit-opinion words.
Design/methodology/approach – This paper proposes a detection algorithm to a medical-opinion-
mining system using rule-based and support vector machines (SVM) algorithms. A corpus from
225 drug reviews were manually annotated by a medical expert for training and testing.
Findings – The results show that SVM significantly outperforms a rule-based algorithm. However,
the results of both algorithms are encouraging and a good foundation for future research. Obviating
the limitations and exploiting combined approaches would improve the results.
Practical implications – An automatic extraction for adverse drug effects information from online
text can help regulatory authorities in rapid information screening and extraction instead of manual
inspection and contributes to the acceleration of medical decision support and safety alert generation.
Originality/value – The results of this study can help database curators in compiling adverse drug
effects databases and researchers to digest the huge amount of textual online information which is
growing rapidly.