Sentiment analysis model to emphasize the impact of online reviews in healthcare industry

Published date14 August 2017
Pages471-486
Date14 August 2017
DOIhttps://doi.org/10.1108/OIR-08-2015-0289
AuthorA.M. Abirami,A. Askarunisa
Subject MatterLibrary & information science,Information behaviour & retrieval,Collection building & management,Bibliometrics,Databases,Information & knowledge management,Information & communications technology,Internet,Records management & preservation,Document management
Sentiment analysis model to
emphasize the impact of online
reviews in healthcare industry
A.M. Abirami
Thiagarajar College of Engineering, Madurai, India, and
A. Askarunisa
Vickram College of Engineering, Sivagangai, India
Abstract
Purpose The purpose of this paper is to develop a systematic approach to extract usersfeelings and
emotions about their experiences in hospitals from online reviews and rank the places using multi-criteria
decision making (MCDM) techniques based on the aggregated sentiment score.
Design/methodology/approach The proposed model used a linguistic approach to extract the sentiment
words from the free text. It used term frequency-inverse document frequency values to represent features of
various places in bag-of-words format. Sentiment dictionary is used to calculate senti-scores. It used different
MCDM techniques like simple additive weight and Technique for Order Preference by Similarity to Ideal
Solution (TOPSIS) methods for ranking hospitals based on their aggregated senti-score.
Findings Statistical correlation analysis between the rankings of places reveals that the TOPSIS method is
the most suitable ranking technique among other MCDM techniques. By improving the senti-score, one can
bring their enterprise to the top position.
Research limitations/implications Data set is collected from different websites like Twitter, Facebook,
etc., for various services/features. Moderate amount of reviews are collected for each place. But not all users
give their views on the social media websites. It would be essential to collect responses from all the customers
who avail different services at different places.
Practical implications The sentiment analysis model proposed in this paper enables B2C and C2C
commerce. Business may take suitable measures to overcome their issues/problems raised by the consumer.
Consumers can share and educate other consumers about their experiences.
Social implications The development of internet has strong influence in all types of industries like
healthcare. The availability of internet has changed the way of accessing the information and sharing their
experience with others. This paper recognizes the use and impact of social media on the healthcare industry
by analyzing the usersfeelings expressed in the form of free text. A suitable decision-making technique is
applied to rank the places, which enables the users to plan their treatment place in a better way.
Originality/value The paper develops a novel approach by applying the TOPSIS method to rank the
different alternative places of the healthcare industry by using the senti-score derived from the usersfeelings,
emotions and experiences expressed in the form of free text.
Keywords Sentiment analysis, TOPSIS method, MCDM technique, Social media analysis
Paper type Research paper
1. Introduction
Most of the healthcare organizations move to electronic medical records and start using
information technology to improve their business intelligence. The amount of data available
to clinicians continues to grow rapidly. With the advent of social media networking sites,
any type of user can share their experiences with others. Nowadays it is common to write in
social media to express ones views. Many organizations laid foundation for the idea that the
healthcare industry can use social media as a tool for the quality improvement by tapping
into the power of online conversations. Social media text related to healthcare has its own
advantages like maintaining personal health diary, augmenting telemedicine, disaster
alerting, etc. These sites help their users to be more cautious about their health and the
quality of service provided by the healthcare industry. Social media, as a data source,
contains valuable consumer insights and enables business intelligence. In the same way,
there has been an increase in attention on social media as a source of research data in areas
Online Information Review
Vol. 41 No. 4, 2017
pp. 471-486
© Emerald PublishingLimited
1468-4527
DOI 10.1108/OIR-08-2015-0289
Received 31 August 2015
Revised 28 January 2016
Accepted 12 May 2016
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/1468-4527.htm
471
Sentiment
analysis model
such as decision making, recommender systems, etc. Much of this research has been
undertaken in the marketing and retail sector to improve customer satisfaction, recommend
new products and so on. But for the healthcare industry, the challenge lies in collecting
medical related data, identifying patterns, creating data models, performing data analysis
and finally using these data to help improve quality of care. It becomes an emerging focus to
monitor its reputation for the purpose of quality improvement.
Opinions are central to almost all human activities and are key influencers of our
behaviors. The beliefs and perceptions of reality, the choices made and the tradition upon
how others view and evaluate the products form the basis for the sentiment analysis.
The enormous gro wth of social medi a through online reviews, discussions, blogs,
micro-blogs, Twitter, etc., on the web enables individuals and organizations to make
decisions using these content. The power of social media is described (Kumar, 2014) and it
can be effectively used to influence public opinion or research behavior. It becomes
common nowadays to ask ones friends and family for opinions before buying any new
product or avail new service. Extracting and summarizing their opinions is called as
opinion mining. In recent years, it has been witnessed that opinionated postings in social
media have helped in business improvement, and have a great impact on social and
political systems (Liu, 2012). British Columbia Safety and Quality Council (2014) report
says that there is improvement in the healthcare industry by engaging patients and
healthcare providers through social media. It recommends many business strategies to
determine the customer insights.
The feature-based sentiment analysis refers to determining the opinions expressed on
different features or aspects of entities like the service of a restaurant, or the picture quality
of a camera. Different features can generate different sentiment responses, for example, the
taste of food may be good but the cost may be high. It involves several sub-processes like
identifying relevant entities, extracting their features, and determining whether the opinion
on the feature is positive, negative or neutral. Although several sentiment analysis
approaches have been proposed for the extraction of the emotional information from
customer reviews, however, the effective analysis on user reviews is still being a research
focus. This paper presents a framework for ranking the hospitals by combining sentiment
analysis and multi-criteria decision making (MCDM) techniques. MCDM methods are
originally developed to solve decision problems with conflicting and non-commensurable
criteria, assuming that compromise is acceptable for conflict resolution, the decision maker
wants a solution that is the closest to the ideal, and the alternatives are evaluated according
to all established criteria (Caterino et al., 2008).
This paper proposes a sentiment analysis model which involves gathering of user
reviews about hospitals in various cities of India from different social media sites.
The model analyzes the sentiments from those reviews and ranks the features of hospitals
using MCDM techniques based on the usersjudgment. The proposed model not only helps
users in decision making, but also in business intelligence for the promotion of healthcare
and gives quality indicators of services related to the industry. This paper is organized as
follows: Section 2 describes the related work, Section 3 explains the proposed methodology,
Sections 4 and 5 present the implementation details and the results and the Section 6
concludes the paper.
2. Related work
Social media has a great impact on all domains and industries. This paper involves the
study of impact of social media on the healthcare industry. A literature review has been
done in three different ways impact of the social media analysis in business or industry,
application of MCDM techniques for industry problems and natural language processing
(NLP) techniques for sentiment analysis.
472
OIR
41,4

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT