=Paper=
{{Paper
|id=Vol-3159/T8-5
|storemode=property
|title=Analyzing COVID-19 Vaccination
|pdfUrl=https://ceur-ws.org/Vol-3159/T8-5.pdf
|volume=Vol-3159
|authors=Poulami Ghosh,Sayani Ghosh
|dblpUrl=https://dblp.org/rec/conf/fire/GhoshG21
}}
==Analyzing COVID-19 Vaccination==
https://ceur-ws.org/Vol-3159/T8-5.pdf
Analyzing COVID-19 Vaccination
POULAMI GHOSH1 , SAYANI GHOSH2
1
Dept. of CA, University of Engineering Management, Kolkata, INDIA
2
Dept. of BCA, University of Engineering Management, Kolkata, INDIA
Abstract
The outbreak of the coronavirus has resulted in unprecedented action, which has led authorities to
decide to begin the blockade of the areas most hit by the infectious disease. Social media has been an
important support for people during this difficult time. On November 9, 2020, when the first vaccine with
an infection rate of 90% or higher was announced, social media responded with , and people around the
world began to express the feelings of vaccination. It was no longer a hypothesis, but closer to ,every day
to become a reality Therefore, it becomes imperative to verify some of the information posted on social
media during the pandemic situation, specially related to Covid vaccines. To this end, it is necessary to
correctly identify fact-checkable posts, so that their information content can be verified.In this work, we
have addressed the problem to identify 3 types of classification on the Twitter microblogging site. We
organized a shared task in the FIRE 2021 conference to study the problem of identifyefficient classifier for
prediction tweets posted during a particular pandemic scenario (the Covid 19). This paper describes the
dataset used in the shared task, and compares the performance of different classification that are provax,
antivax and last neutraal for identifying effective tweets related to Covid vaccines.We experimented
with a classification-based approach. Our experiment shows that SVM classification performs well in
order to effiective posts.Using this support vector machine in order to solve the antivax, provax,neutral
classification of twets .We’re going to do this because vaccination is an important step for Covid19 so
people can easily fix the news about the vaccine and grab their own slot.
Keywords
Corona, Twitter, Microblogs, vaccine, classification, classifier
1. Introduction
The use of social media is escalating worldwide as lockdown is in place in some parts of the
world and social distancing in other parts of the world . To exchange ideas and information
regarding a series of aspects that occurred during this period. People also seem to be relying on
information posted on social media. As a result, the social media platform is gaining increasing
attention as it is a moderator channel between each individual and another of the world, and has
become one of the fastest growing information systems for social applications. In this channel,
individuals express different views, opinions, and feelings during various events triggered by the
coronavirus pandemic.The coronavirus outbreak caused by the novel coronavirus SARSCoV2
has resulted in a series of changes in many aspects of the economic and social lives of many
people. Since its onset, the coronavirus pandemic has continued to monopolize different regions
of the world, reaching 220 countries and territories on December 9, 2020[1]. Among several
Forum for Information Retrieval Evaluation, December 13-17, 2021, India
" poulami.ghosh@uem.edu.in (P. GHOSH); sayanighosh37@gmail.com (S. GHOSH)
© 2021 Forum for Information Retrieval Evaluation, December 13-17, 2021, India
CEUR Workshop Proceedings (CEUR-WS.org)
well-known social media platforms, Twitter has gained particular attention because users can
easily disseminate information about their opinions on a certain topic via a public message,
called a tweet[2]. In addition to the information provided voluntarily by the user, a tweet may
also contain information relating to the location of user and may contain links, emojis, and
hashtags that may help users to better express their feelings. , which makes it an excellent
source of valuable information [2]. Even more, Twitter has been used by government officials
and politicians to inform the public about their activities or, in case , major events occurring
. We introduced the problem to identify three classification on tweets as a shared task titled
‘Information Retrieval from Microblogs During Disaster’ (IRMiDis). Collects and annotates
the COVID19 vaccination data set, determines the COVID19 vaccination posture detection
classifier, and the number of reported events and tweets and posture (e.g., in favor, against
or neutral) linked to There is. years of analysis by the media.Many teams participated in the
IRMiDis shared task, and proposed several classifier to idenfy.The methodology was based on
classification of tweets into three classes, viz. provax) and antivax and neutral using Support
Vector Machine (SVM) model that is known to perform well for text classification. We have
used SVM here because it is a very efficient simple classifier algorithm which is widely used
for pattern recognition which can also have a very good classification performance than any
other classifier.
2. RELATED WORK
There has been a lot of research in recent years on utilizing online social media (OSM) during
disasters, which involves several challenges such as, parsing short and informal messages,
handling information overload, and prioritizing different types of information. The reader is
referred to [1],[3] for comprehensive surveys on using OSM for disaster informatics. Several
studies have attempted to identify particular types of information from microblogs posted during
disaster events. For instance, some studies focused on extraction of situational information
[4],[1]or actionable posts[3], while some works attempted to identify more specific information
such as need of resources or availability of resources . Many of these prior works use either
classification-based approaches (e.g., classifying tweets into situational and non-situational
classes [4]. Machine learning methods include classic machine learning and deep learning
algorithms. Classical machine learning algorithms frequently used for posture detection are
support vector machines (SVMs) and many types[5] In the present work, we focus on identifying
effective classifier for classification on tweets data.
3. MICROBLOG DATASET
The present work is a task (IRMiDis) that we organized at the Annual Conference of the Forum
for Information Retrieval Evaluation (FIRE) 2021. The task was to identify antivax,provax,neutral
tweets that is related to covid vaccine from among a large set of tweets posted during Covid 19.
This section describes the dataset (that was used in the shared task).
3.1. Collecting tweets during Covid 19
We used Twitter search API2 to crawl 100k English tweets related to Covid-19 Vaccine which
have been occurring worldwide since November-December 2019 with the keywords "covid",
"pandemic’, and ’vaccine’. We then removed the dataset redundancy and nearly duplicate
tweets to get a chronologically sorted set of unique English tweets (based on Twitter-assigned
timestamps).
4. PROPOSED METHOD
The design of our proposed system for classification of tweets into provax,antivax,neutral is
shown in the Fig. 2. The steps of designed model are provided as follows.
5. METHODOLOGY
The main objective of this work was to classify the tweets into provax, antivax and neutral
checkable accurately from a set of tweets that were gathered. In order to fulfill our purpose,
our proposed framework first collects the set of tweets and then Support Vector Machine
classification was applied in order to recognize the provax,antivax, neutral tweets properly. Our
proposed methodologies are ‘semi-automatic’ in nature, where some manual effort is employed
to generate a training set for the classifier, and then the classification and ranking are automatic.
5.1. Pre-processing the tweets
In the current study 2792 tweets have been used to revel the antivax/provax/neutral sentiment
from a tweet. The raw tweets are preprocess ,and in this steps we clean data, remove stop words,
white spaces then the tweets are case-folded to lowercase.
5.2. Constructing training set for classification
After streaming and lemmatization are applied,the pre processed level tweets are then transform
to numeric feature vector using term frequency inverse document frequency. After transforming
the unstructured tweet data into numeric structured data Then classifier SVM (described later)
is trained over this training set containing 991 provax tweets and 791 antivax tweets and 1010
tweets are neutral and get training accuracy 0.67.
5.3. Support Vector machine
Support vector machines (SVMs) are a family of supervised learning algorithms used for classi-
fication, regression , and other tasks such as outlier detection.Other classification algorithms
suffer from over-compliance, but one of the advantages of SVM is that this situation will be
difficult . Another advantage lies in the fact that, in addition to binary classification, multi-class
classification can also be performed by by combining several binary classification functions.
Thus, each class in finds a classifier that separates it from the other classes for each class con-
sidered individually at a time.The SVM algorithm builds an N-dimensional hyperplane model
precision recall fi-score support
0 0.71 0.57 0.63 209
1 0.66 0.77 0.71 246
2 0.65 0.65 0.65 243
accuracy 0.67 699
macro avg 0.67 0.66 0.67 698
micro avg 0.67 0.67 0.67 698
Table 1
classification result
that assigns future instances to one of two possible output classes.In our task we are applied
SVM to solve the various type of classification of tweets.
5.4. Identify fact-checkable tweets
The training dataset was compared with the trained dataset to identify three type of classification
provax,antivax, and neutral tweets.
6. RESULTS AND DISCUSSION
In this task we firstly Preprocess the tweets after that converts the tweets into a feature vector
then transforming the unstructured tweet data into numeric structured data the same has been
used to train a SVM in order to solve the antivax,provax, neutral classification of tweets and at
last, after a set of test and set of sorters, we come to the conclusion that the accuracy of the
final result is 0.380 and the macro F1 score is 0.370. The result is shown in table 1.
7. CONCLUSION
The purpose of this treatise was to monitor the number of Covid-19 vaccine twitter messages
during the analysis period, matching major events reported by tweets for Covid-19 vaccination.
The proposed approach categorized tweets. Covid-19 is divided into three major classes, Provax,
Antivax and Neutral, in relation to vaccination. We educated the system on intermediate datasets
to improve the accuracy of the system and evaluated the system on challenging datasets to prove
robustness and durability. System reliability. It can be said that the system can be implemented
with a dataset of tweets incidental to the recommendation and related large tweets in the future.
And again, this operation may be further extended using other sorters to see the changes in the
results.
References
[1] Worldometer, Coronavirus update (live): 63,777,845 cases and 1,477,777 deaths from covid-19
virus pandemic., 2020. URL: https://www.worldometers.info/coronavirus/.
[2] E. D’Andrea, P. Ducange, A. Bechini, A. Renda, F. Marcelloni, Monitoring the public opinion
about the vaccination topic from tweets analysis, Expert Systems with Applications 116
(2019) 209–226. URL: https://www.sciencedirect.com/science/article/pii/S0957417418305803.
doi:https://doi.org/10.1016/j.eswa.2018.09.009.
[3] K. Rudra, N. Ganguly, P. Goyal, S. Ghosh, Extracting and summarizing situational infor-
mation from the twitter social media during disasters, ACM Trans. Web 12 (2018). URL:
https://doi.org/10.1145/3178541. doi:10.1145/3178541.
[4] K. Rudra, S. Ghosh, N. Ganguly, P. Goyal, S. Ghosh, Extracting situational information
from microblogs during disaster events: A classification-summarization approach, in:
Proceedings of the 24th ACM International on Conference on Information and Knowledge
Management, CIKM ’15, Association for Computing Machinery, New York, NY, USA, 2015, p.
583–592. URL: https://doi.org/10.1145/2806416.2806485. doi:10.1145/2806416.2806485.
[5] M. Basu, S. Ghosh, K. Ghosh, Overview of the fire 2018 track: Information retrieval
from microblogs during disasters (irmidis), in: Proceedings of the 10th Annual Meeting
of the Forum for Information Retrieval Evaluation, FIRE’18, Association for Computing
Machinery, New York, NY, USA, 2018, p. 1–5. URL: https://doi.org/10.1145/3293339.3293340.
doi:10.1145/3293339.3293340.
Set of Tweets
Collected data
Dataset clearing
Classifier
Identify
antivax,provax, neutral
tweets
Figure 1: Proposed system for Classification of antivax, provax, neutral tweets