A deep neural networks-based fusion model for COVID-19 rumor detection from online social media
| DOI | https://doi.org/10.1108/DTA-06-2021-0160 |
| Published date | 22 April 2022 |
| Date | 22 April 2022 |
| Pages | 806-824 |
| Subject Matter | Library & information science,Librarianship/library management,Library technology,Information behaviour & retrieval,Metadata,Information & knowledge management,Information & communications technology,Internet |
| Author | Heng-yang Lu,Jun Yang,Wei Fang,Xiaoning Song,Chongjun Wang |
A deep neural networks-based
fusion model for COVID-19 rumor
detection from online social media
Heng-yang Lu
Jiangsu Provincial Engineering Laboratory of Pattern Recognition and
Computational Intelligence, Jiangnan University, Wuxi, China;
School of Internet of Things Engineering, Jiangnan University, Wuxi, China and
State Key Laboratory for Novel Software Technology, Nanjing University,
Nanjing, China
Jun Yang
Marcpoint Co., Ltd., Shanghai, China
Wei Fang and Xiaoning Song
Jiangsu Provincial Engineering Laboratory of Pattern Recognition and
Computational Intelligence, Jiangnan University, Wuxi, China, and
Chongjun Wang
State Key Laboratory for Novel Software Technology, Nanjing University,
Nanjing, China
Abstract
Purpose –The COVID-19 has become a global pandemic, which has caused large number of deaths and
huge economic losses. These losses are not only caused by the virus but also by the related rumors.
Nowadays, online social media are quite popular, where billions of people express their opinions and
propagate information.Rumors about COVID-19 posted on online social media usually spread rapidly; it is
hard to analyze and detect rumors only by artificial processing. The purpose of this paper is to propose a
novel model called the Topic-Comment-based Rumor Detection model (TopCom) to detect rumors as soon
as possible.
Design/methodology/approach –The authors conducted COVID-19 rumor detection from Sina Weibo, one
of the most widely used Chinese online social media. The authors constructed a dataset about COVID-19 from
January 1 to June 30, 2020 with a web crawler, including both rumor and non-rumors. The rumor detection task
is regarded as a binary classification problem. The proposed TopCom model exploits the topical memory
networks to fuse latent topic information with original microblogs, which solves the sparsity problems brought
by short-text microblogs. In addition, TopCom fuses comments with corresponding microblogs to further
improve the performance.
Findings –Experimental results on a publicly available dataset and the proposed COVID dataset have shown
superiority and efficiency compared with baselines. The authors further randomly selected microblogs posted
from July 1–31, 2020 for the case study, which also shows the effectiveness and application prospects for
detecting rumors about COVID-19 automatically.
Originality/value –The originality of TopCom lies in the fusion of latent topic information of original
microblogs and corresponding comments with DNNs-based models for the COVID-19 rumor detection task,
whose value is to help detect rumors automatically in a short time.
Keywords Rumor detection, COVID-19, Deep neural networks, Data analysis, Document characterization,
Information fusion
Paper type Research paper
DTA
56,5
806
Funding: This research was funded by the National Key Research and Development Program of China
[No. 2020YFA0908300], the National Natural Science Foundation of China [Grant No. 62002137,
61876072], the Fundamental Research Funds for the Central Universities [No. JUSRP12021], and the
State Key Lab. for Novel Software Technology, Nanjing University, P.R. China [No. KFKT2020B02].
The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/2514-9288.htm
Received 18 June 2021
Revised 15 January 2022
31 January 2022
Accepted 2 April 2022
Data Technologies and
Applications
Vol. 56 No. 5, 2022
pp. 806-824
© Emerald Publishing Limited
2514-9288
DOI 10.1108/DTA-06-2021-0160
1. Introduction
The new coronavirus disease 2019 (COVID-19) has become a global pandemic and keeps
breaking out worldwide. This terrible pandemic has caused a lot of deaths and huge economic
losses. Among them, we have witnessed that rumors about COVID-19 keep appearing on
online social media (Shi et al., 2020;Sharma and Kapoor, 2022;Wang et al., 2021). For example,
the end times conspiracy theorist Rick Wiles declared that COVID-19 vaccines are a plot to
depopulate the world by changing DNA. The rapid spread of this kind of rumor may block
the vaccination and cause huge losses in various domains. So, it is quite important to detect
and dispel rumors about COVID-19 accurately as soon as possible.
There are various opinions upon rumors. A very early literature claimed a rumor is a
widely circulated proposition related to current events, but not officially confirmed, with the
purpose of making people believe (Knapp, 1944). Shibutani believed that rumors are
improvised news generated during the discussion of a group of people (Shibutani, 1966).
From the opinions of Peng, rumors are widespread, unproven information in a dangerous or
potentially threatening situation (Peng et al., 2018).
Online social media naturally provides the platform for people to discuss various events,
which may producerumors. Discussions about certainevents or topics on online socialmedia
are usually in the formats of conversations, including a source post, corresponding replies
(comments)and reposts. Researches of rumordetection at a macro level aimto decide whether
the conversationsrelating toa certain event belong to arumor or not by analyzing surrounding
discourses (Procter et al., 2013;Wu et al.,2015;Zubiaga et al., 2016). This is similar to the
supporting, denying, querying or commenting (SDQC) rumor detection task introduced in
SemEval workshop (Derczynski et al.,2017;Gorr ell et al., 2019). Another kind of micro-level
rumordetection aims to detect whethera single post belongs to a rumor or not.This is similar to
the veracity prediction rumor detection task, which focuses on the credibility of single posts
(Sicilia et al., 2017,2018). Both tasks can be modeled as theclassification problem. One of the
commonresearch lines belongsto designing feature-basedmethods. The main ideais manually
discoveringtext features from contents(Qazvinian et al., 2011;Popat,2017;Castillo et al., 2013).
This kind of method highly depends on the hand-crafted features, which is limited to prior
knowledge and time costs. With the development of deep learning, it is a common way to
automaticallyextract high-dimensional semanticfeatures with deep neural networks (DNNs).
For example,the latest researches exploit DNNslike Gated recurrent unit (GRU) networksand
convolutional neural networks (CNNs) to detect rumors (Ma et al., 2016;Yu et al.,2017;Bian
et al.,2020). Because of the effectiveness of automatic feature extraction, these DNNs-based
methodshave become more and more popular recently.Another research linemakes full use of
online social information for improving the detectionof rumors, such as propagation patterns
and interactions (Maet a l., 2018;Shu et al., 2018;Kwon et al., 2013).
This paper aims to detect rumors about COVID-19 from microblogs posted on Sina Weibo,
which is a popular Chinese online social media platform. Figure 1 simply demonstrates a
rumor posted on Sina Weibo in Chinese.
Users usually post microblogs on Sina Weibo to express their opinions and share information.
Different from online news, which is also one of themain sources of rumors, microblogsfrom online
social media are usually short in length, which may cause the sparsity problem for the lack of
words. Furthermore, we can often find additional social information such as comments along with
corresponding microblogs. These comments can bring a positive influence to detect rumors. For
example, in Figure 1,thecommentonthefirstlineclaims“Fake? Is there a lotus in the sea?”.The
comment on the second line claims that “itis by photoshop”. These comments have shown their
strong doubts about the authenticity of the proposed microblog, which can help make judgments.
It is necessary to detect COVID-19 rumors from online social media by considering both
microblogs and corresponding comments. This paper aims to detect rumors about COVID-19
with AI techniques from microblogs posted on Sina Weibo. The main contributions include:
TopCom model
for COVID-19
rumor
detection
807
To continue reading
Request your trial