Online investigation of users’ attitudes using automatic question answering

DOIhttps://doi.org/10.1108/OIR-10-2016-0299
Published date11 June 2018
Pages419-435
Date11 June 2018
AuthorChengzhi Zhang,Qingqing Zhou
Subject MatterLibrary & information science,Information behaviour & retrieval,Collection building & management,Bibliometrics,Databases,Information & knowledge management,Information & communications technology,Internet,Records management & preservation,Document management
Online investigation of users
attitudes using automatic
question answering
Chengzhi Zhang and Qingqing Zhou
Department of Information Management,
Nanjing University of Science and Technology, Nanjing, China
Abstract
Purpose With the development of the internet, huge numbers of reviews are generated, disseminated, and
shared on e-commerce and social media websites by internet users. These reviews usually indicate users
opinions about products or services directly, and are thus valuable for efficient marketing. The purpose of
this paper is to mine online usersattitudes from a huge pool of reviews via automatic question answering.
Design/methodology/approach The authors make use of online reviews to complete an online
investigation via automatic question answering (AQA). In the process of AQA, question generation and
extraction of corresponding answers are conducted via sentiment computing. In order to verify the
performance of AQA for online investigation, online reviews from a well-knowntravel website, namely Tuniu.
com, are used as the experimental data set. Finally, the experimental results from AQA vs a traditional
questionnaire are compared.
Findings The experimental results show that results between the AQA-based automatic questionnaire and
the traditional questionnaire are consistent. Hence, the AQA method is reliable in identifying usersattitudes.
Although this paper takes Chinese tourism reviews as the experimental data, the method is domain and
language independent.
Originality/value To the best of the authorsknowledge, this is the first study to use the AQA method to
mine usersattitudes towards tourism services. Using online reviews may overcome problems with using
traditional questionnaires, such as high costs and long cycle for questionnaire design and answering.
Keywords Sentiment analysis, Automatic question answering, Review mining, User survey
Paper type Research paper
1. Introduction
Usersattitudes towards services or products have become a hot topic in both academia and
industry. Attitudes play an important role in product design, marketing, etc. Users can make
purchase decisions quickly according to attitudes of other users, while enterprises can
improve the quality of products and services and develop marketing programs effectively.
Most researchers have analyzed usersattitudes by means of questionnaires, which have
solid theories and massive practices (Hayes, 2008; Mochimaru et al., 2012), but are expensive
and time-consuming. How can we overcome these shortcomings?
Web 2.0 has enabled to development of e-commerce and social media, which in turn
attracts huge numbers of users. A report from We Are Social[1] showed that by January
2016, the number of active social media users globally was 2.3 billion. These users
generate huge numbers of online reviews, which express their attitudes towards product
performances and service quality. Figure 1 shows two examples of online reviews,
wherein both users express attitudes about their comment targets. Hence, by assessing
mass user reviews, investigations into usersattitudes can be conducted automatically,
which may overcome shortcomings of traditional questionnaires. For example, Kramer
et al. (2014) used 689,003 Facebook users to test massive-scale emotional contagion
through social networks. Compared with traditional questionnaires, Kramer et al. (2014)
have a much larger sample size and a lower cost.
How can we identify usersattitudes effectively? In this paper, automatic question
answering (AQA) is conducted on online reviews from a travel website: Tuniu.com[2] to
Online Information Review
Vol. 42 No. 3, 2018
pp. 419-435
© Emerald PublishingLimited
1468-4527
DOI 10.1108/OIR-10-2016-0299
Received 8 October 2016
Revised 14 September 2017
Accepted 27 October 2017
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/1468-4527.htm
419
Online
investigation
automatically analyze usersattitudes. In the process of applying AQA, questions are
generated and corresponding answers are extracted using sentiment analysis. Specifically,
the paper generates questions based on question templates, and extracts aspects as question
options via a word pair method; extracts corresponding answers using sentiment lexicon-
based aspect-level sentiment analysis; and compares the results with those obtained via
traditional questionnaires. The experimental results show that usersattitudes as analyzed
by AQA method are consistent with those extracted from traditional questionnaires.
Hence, the AQA method is reliable in mining usersattitudes. In addition, the method is
domain and language independent.
The remainder of this paper is organized as follows. In Section 2, related work is
reviewed. Methodology is introduced in Section 3. Data collection method and comparative
analysis results are presented in Section 4. Section 5 discusses AQA-based automatic
questionnaires vs traditional questionnaires for mining usersattitudes. Section 6 provides
the conclusion and suggestions for future works.
2. Literature review
Traditional methods such as questionnaire, observation, and face-to-face interviews have
been widely employed for analyzing usersattitudes towards services and products (Beiske,
2002; Borelli, 2014; Phellas et al., 2011). However, the sheer effort required from surveyors,
and the cost of resources, are relatively high. This paper uses the AQA method to analyze
usersattitudes towards tourism services with online reviews. Hence, there are three areas of
research related to this study: online user surveys via e-commerce and social media
websites; AQA; and review mining and sentiment analysis.
2.1 Online user survey via e-commerce and social media websites
With the development of Web 2.0, many researchers have conducted user surveys via social
media and e-commerce websites. For example, Lukas (2008) conducted an open-ended
qualitative survey about the choices people made when choosing profile pictures on
Facebook. The results demonstrated that women tended to change their profile image more
often. Liu et al. (2015) analyzed userslife satisfaction by mining usersFacebook status, and
proved that user-generated content reflected userspsychological states. Settanni and
Marengo (2014) supported the feasibility and validity of studying individual emotional
well-being by examining Facebook profiles. Qiu et al. (2015) considered the association
between selfies and personality by measuring participantspersonality traits and coding
their selfies posted on social networking sites. Brandt (2012) used social media as a tool in
marketing research, and examined whether website ratings were similar to ratings captured
through the traditional survey method. Ruizmafe et al. (2014) identified the main drivers of
Facebook fan page loyalty so as to promote the creation of affective links and long-term
relationships with users. Pasternak et al. (2015) explored the nature of consumer
participation in eWOM activities on Facebook brand pages. Vidal et al. (2015) used 69,961
tweets to investigate food-related consumer behaviors, and found that Twitter data merits
inclusion in the researchers toolbox. He et al. (2016) explored how to use social media in
e-government to strengthen interactivity between government and the general public.
A Very Useful Book
Well written, well laid out and (best of all)
an exceedingly useful treatment of machine
learning and predictive data analytics. Highly
recommended
Interesting
This is awesome. Worth the trip. I think it
is the largest reclining Buddha in the world
but I am not sure
(a) A book review from Amazon.com (b) A tourism review from tripadvisor.com
Figure 1.
Examples of online
reviews
420
OIR
42,3

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