Volume 8, Issue 4, April – 2023 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

An Analytical Insight of Omicron Sentiments by

N-Gram Using Machine Learning
N. Narasimha Rao1 V.Srujan2
1 2
Assistant Professor, NRI Institute of Technology, UG Scholar, Dept. of IT, NRI Institute of Technology,
A.P, India-521212 A.P, India-521212

A. Praneeth Surya3 D. Siva Teja4

3 4
UG Scholar, Dept. of IT, NRI Institute of Technology, UG Scholar, Dept. of IT, NRI Institute of Technology,
A.P, India-521212 A.P, India-521212

Abstract:- The capacity to assess and forecast a variety evaluating public sentiment using data science techniques
of topics, including commercial requirements, including natural language processing and machine learning
environmental needs, election patterns (polls), approaches. Twelve well-known machine learning
governmental needs, etc., may be added to social media algorithms are utilised in the suggested research paper to
as an intelligent platform. This inspired us to start a analyse public opinions. Commonly used words are
thorough investigation of public thoughts and opinions represented as n-grams; three of these n-grams—Unigram,
on the COVID-19 epidemic on Twitter. The fundamental Bigram, and Trigram—are gathered here, and predictions
training data were gathered from tweets. Based on this, are made using the data .Today's online media has
we have produced research using ensemble deep developed a reputation for its ability to switch as well as
learning algorithms to forecast Twitter views more advertise. People divide their pricey opinions, assessments,
accurately than earlier works that do the same task. An and experiences on responsive destinations with the hope
N-gram stacked auto encoder supervised learning that others would profit from these. Twitter is one of these
technique is used to extract features first. The collected platforms where the general public communicates its
features are subsequently used in a classification and opinions in brief terms, like 140 characters. Twitter serves
prediction process using an ensemble fusion strategy as the corpus for open mining and sentiment analysis. These
comprising certain machine learning algorithms, audits continue to be for anything and everything other than
including decision trees (DT), support vector machines management, including movies, financial transactions,
(SVM), random forests (RF), and K-nearest neighbors educational institutions, legal matters, and a great deal more.
(KNN). Using both mean and mode approaches, all People provide their unbiased opinions about anything they
individual findings are combined/fused for a superior wish in order for this audit to be seen as more
forecast. The N-gram stacking encoder we suggest using comprehensive and real.
in combination with an ensemble machine learning
strategy surpasses all other known competitive To complete this entire framework, five basic
techniques, including bigram auto encoders and unigram advancements are necessary. The first step is choosing how
auto encoders. The public has a great deal of trust in to prepare the data based on the type of concern. The second
government policy during the third wave, and they step is preprocessing the data to remove irrelevant
support all measures taken to contain the epidemic, information such as URLs, customer names, shoptalk
including widespread participation in vaccine vocabulary, imagery, and so on. [fig:7.1]. The third step is to
programmes.. The study's findings may be summarised establish associations through Twitter knowledge
by saying that people are getting past their fear of the computation. Naive Bayes and Support Vector Machine are
disease. used for the alliance of tweets interested in different classes.
The final step is to reveal the advance results.
Keywords:- Omicron Sentiment Analysis, N-Gram, Analysis,
Social Media, Omicron, Tweets, Twitter, Big Data, Data II. TECHNOLOGIES USED
Analysis.

I. INTRODUCTION

Internet users have been growing quickly over the past

ten years, and with the Covid-19 epidemic, social media
became the preferred medium for expressing public
sentiment. They are utilising the free microblogging website
Twitter to impulsively share their ideas, happiness, and
sadness. In order to forecast public opinions on socially
relevant problems, researchers are very interested in Fig 1 Technologies Used

IJISRT23APR1751 www.ijisrt.com 1441

Volume 8, Issue 4, April – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
 KNN
KNN is a non-parametric, slow learning method. In
order to forecast the categorization of the next sample point,
it leverages data from many classes. KNN is non-parametric
since the model is distributed from the data and no
assumptions are made about the data being
investigated.[Fig:2]

Fig 4 SVM

 N-GRAM:
I believe that N-gram is the simplest concept to
comprehend in the entire field of machine learning. A
combination of N words in a row is called an N-gram. For
illustration, "Medium blog" is a two-word combination (a
bigram), "A Medium blog post" consists of four words (a 4-
gram), and "Write on Medium" has three words (trigram).
Fig 2 KNN That was quite dull and uninspiring. Indeed, yet we still
have to take into account the likelihood associated with n-
 Random Forest grams, which is quite intriguing.[Fig:5]
A popular classification and regression approach is
Random Forest. We may claim that the Random Forest
Algorithm is one of the most significant algorithms in
machine learning since classification and regression are the
most significant parts of machine learning. The ability to
categorize observations accurately is useful for a variety of
commercial applications, such as determining whether a
certain user will purchase a product or if a loan would fail or
not.[Fig:3]

Fig 5 N-GRAM

III. SOFTWARE REQUIREMENTS

SPECIFICATION

SRS is a captures complete description about how the

system is expected to perform. It is usually signed off at the
end of requirements engineering phase . It defines how
software system will interact with all internal modules,
hardware, communication with each other programs and
human user interactions with a wide range of real like
scenarios.

Fig 3 Random Forest

 SVM:
SVMs could offer a learning technique that is
applicable to both regression and classification. A fast
algorithm that produces favorable outcomes for a multitude
of educational assignments is classified. It is not based on
probability. A binary linear classifier that takes a set of input
data and predicts, for every given input, which of the two
available outcomes it belongs to.Classes are made up of the
input. The support vector is composed of the training
examples that are used for its formation. apparatus.[Fig:4] Fig 6 SRS

IJISRT23APR1751 www.ijisrt.com 1442

Volume 8, Issue 4, April – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
 Reliability  Advantages of Proposed System:
It would be more dependable and maintain the web
application's updated information current. Once logged in,  Yields rich insights as it include qualitative and
our username and password are hidden from view and are quantitative analysis.
not visible to other users of the online application.  Easily replicable since it follows systematic procedures.
 Relatively inexpensive.
 Quality
This project has higher quality, and students, VI. SYSTEM ARCHITECTURE
instructors, and administrators may access it from anywhere
via the internet.

 Maintainability
The administrator would cleanly maintain the
programme to keep the data secure and error-free.

 Efficiency
Downloading the information and answering questions
would be more efficient for students, and instructors may
upload the data as well.

 Portability
It would run without cost in any browser on any
platform.

 Performance Fig 7 System Architecture

Performance is higher because it would have provided
excellent service to both instructors and students. The goal of sentiment analysis is to determine
automatically whether or not a particular textual item
IV. EXISTING SYSTEM expresses opinions like positive or negative on an important
issue. The valuable information from the text is retrieved
We have seen evolution in covid-19 during the last using latent semantic analysis (LSA). Offline sentiment and
two years. There are several varieties of covid, including semantic models have been developed for analysis in order
omicron and delta. Omicron analysis only uses information to assess machine learning methods and get the best
or data that has been personally collected or observed by answers. The K closest neighbour (KNN), Random Forest
humans. Omicron analysis was manually updated. We (RF), Decision Tree (DT), and Support Vector Machine AI
employed machine learning with the n-gram approach to get models have been utilised (SVM). In the proposed study, we
around this. used ensemble learning, combining the predictive models
from the DT, SVM, RF, and KNN, and then combining all
 Disadvantages of Existing System: of the predictions using statistics like the mode or mean to
get better outcomes.
 Time Consuming Process.
 Disregards the Overall Context of the Text .  Data Collection:
Twitter data was gathered for this study using open-
 Can be Reductive in its Approach.
source data from the IEEE website [32]. This publicly
accessible dataset included tweets from across the world that
V. PROPOSED SYSTEM
had been filtered using the terms "coronavirus," "_covid," "-
covid-19," "sarscov2," "#covid19," "#covid 19," "2019-
To analyse the omicron data, we employ machine
ncov," "#2019ncov," "sarscov2," "#covid," "sars cov2," etc.
learning using Python. Using Python, do Omicron Sentiment
Tweet IDs were only accessible as of March 20, 2020 [28].
Analysis When people were discussing the Omicron version
According to the IEEE website, tweet objects are what are
on Twitter, the dataset that we are utilising for the job of
used to collect information about tweets and extract the
Omicron sentiment analysis was first gathered. It may be
tweet ID.[Fig:7]
acquired from Kaggle. So let's begin by importing the
required Python modules and the dataset in order to do the
 Data Preprocessing:
Omicron sentiment analysis operation. All other known
A key component of the social media network idea
competitive approaches, such as the bigram autoencoder and
analysis system is pre-processing the data. That is in the
the unigram autoencoder, are outperformed by our
latent semantic analysis and sentiment analysis of Twitter's
suggested strategy, which incorporates an N-gram stacking
streaming data. The text data that may be accessed via
encoder into an ensemble machine learning scheme.
Twitter are largely jumbled and loud. Data preparation is
essential to provide the best outcome. [Fig:8].

IJISRT23APR1751 www.ijisrt.com 1443

Volume 8, Issue 4, April – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
 Feature Extraction:
Feature extraction in the analysis of textual data is
difficult. Text feature extraction is a method of representing
a text message by removing text information from a vast
volume of text processing. The procedure of feature
extraction, which involves lowering the feature space
dimensions, has been successfully implemented. We
eliminate uncorrelated features during feature extraction. To
recognise sentiment and perform latent sematic analysis at
the word level, we provide an unique deep learning
approach in this proposed study employing stacked
encoders. The distributed word vector representation from
the "n" gram is used as input in the newly suggested model,
and the resulting continuous word vectors are merged with
stacked auto encoder for fine-tuning word embeddings.
Fig 8 Data Preprocessing
 Stacked Autoencoder:
 Data Cleaning:
The following procedures are used to eliminate  Words that are often used are spread in n-gram string
undesirable material during this phase: data.
 The representation is then transformed into a reduced
 Removal of HTML Characters: vector using the "SA" technique of the attacked
HTML elements like >& that are contained in the autoencoders.
original data are frequently seen in online data. We can  With machine learning techniques like decision tree
transform these things into typical HTML tags by utilising (DT), support vector machine (SVM), random forest
Python's HTML parser. (RF), and K-nearest neighbour (KNN), sentiment
analysis and latent semantic analysis are applied
 Punctuation Elimination: (KNN).
All punctuation that is in accordance with the priority  The quick forecast is produced by applying the
must be removed. Examples of necessary punctuation that ensemble approach to the ML model discussed above.
must be kept in place are ".", ",", and "?" Other punctuation The feature extraction framework is depicted in Figure.
marks must be dropped.
VII. FUTURE SCOPE
 Expressions that have been Removed:
The text data may also include human expressions. This study has a number of restrictions. In order to
These phrases should be deleted since they often don't relate find trends and the frequency of keywords used in this
to the text's subject matter. study, keywords connected to Omicron are worth
considering. It's possible that the list of chosen keywords
 Tokenization: isn't full. Sentiment analysis can have an impact on future
Tokenization is the process of dividing lengthy words research in information systems, public mental health, and
or strings into smaller units called tokens. There are two policy formulation. This study provides a helpful analysis to
steps to it. pinpoint the characteristics that lead to posting both good
and negative tweets. We cannot claim that social media is
 Split Attached Words: only to blame for societal responses and the emotional
The initial phase is creating text data in a loose impact it has on individuals. This is just based on
structure in the social network. The majority of tweets correlational study; there are many more aspects that
include phrases like "its epidemic," "totally lockdown day," connect to psychological effect.
and similar expressions. These things can be divided into
their regular forms. VIII. CONCLUSION

 Standardizing Words: In our upcoming research, we will be utilizing more

The textual information is not formatted properly, sophisticated deep learning techniques and multiple
e.g., "loveeee u," "miss u," etc. These phrases need to be classifiers to further enhance the precision (e.g., surpassing
divided into their correct sections. 90%) of sentiment analysis on social mediaposts related to
covid19. This is the method of evaluating the emotions
 Stemming: Related to the Omicron form of coronavirus. The World
The words are being returned to their original form Health Organization has labeled this novel form of
through this procedure. That reduces the amount of words in coronavirus as a variant of concern. I trust that you enjoyed
the text from the root to the word type. As an illustrative reading this piece on sentiment analysis of Omicron using
example, the words "Jumping" and "jumped" will be machine learning techniques.
removed in favour of the word "jump."

IJISRT23APR1751 www.ijisrt.com 1444

Volume 8, Issue 4, April – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
REFERENCES: [14]. Pinter, G.; Felde, I.; Mosavi, A.; Ghamisi, P.;
Gloaguen, R. COVID-19 pandemic prediction for
[1]. M. Lenzerini, “Data integration: A theoretical Hungary; a hybrid machine learning
perspective,” in PODS, 2002, pp. 233– 246. approach. Mathematics 2020, 8, 890. [Google
[2]. D. Caruso, “Bringing Agility to Business Scholar] [CrossRef]
Intelligence,” February 2011, Information [15]. Twitter: Standard Search Api. 2020. Available
Management, online: https://developer.twitter.com/ en/docs/
http://www.informationmanagement.com/nfodirect/2 tweets/ search/ overview (accessed on 20 April
009191/business intelligence metadata analytics ETL 2020).
data management-10019747-1.html.
[3]. R. Hughes, Agile Data Warehousing: Delivering BIOGRAPHIES
world-class business intelligence systems using
Scrum and XP. IUniverse, 2008.
[4]. Y. Chen, S. Alspaugh, and R. Katz, “Interactive
analytical processing in big data systems: A cross-
industry study of map reduces workloads,”
Proceedings of the VLDB Endowment, vol. 5, no. 12,
pp. 1802–1813, 2012.
[5]. M. Singh, A.K. Jakhar, S Pandey Sentiment analysis
on the impact of coronavirus in social life using the
BERT model Mr. N. Narasimha Rao is currently working as
[6]. T. Vijay, A. Chawla, B. Dhanka, P. Karmakar assistant Professor in Information technology at NRI
Sentiment Analysis on COVID-19 Twitter Data 2020 Institute of technology, Pothavarappadu, Agiripalli,
5th IEEE International Conference on Recent Krishna(dist), India. Completed B.tech in Lakireddy Bali
Advances and Innovations in Engineering (ICRAIE) reddy college of engineering and M.Tech in Vikas
[7]. G. Matošević, V. Bevanda Sentiment analysis of group of institutioins
tweets about COVID-19 disease during pandemic
2020 43rd International Convention on Information,
Communication and Electronic Technology (MIPRO)
[8]. xxx Apoorv Agarwal BoyiXie Ilia Vovsha Owen
Rambow Rebecca Passonneau, “Sentiment Analysis
of Twitter Data”, Columbia University, Newyork.
[9]. K.-W. Fu, H. Liang, N. Saroha, Z. T. H. Tse, P. Ip, I.
C.-H. Fung, How people react to Zika virus outbreaks
on Twitter? A computational content analysis, Am. J. V.Srujan is currently studying B. Tech with
Infect. Control 44 (2016) 1700–1702. specification of Information Technology in NRI Institute of
[10]. D. Thorpe Huerta, J. B. Hawkins, J. S. Brownstein, Technology. He done a summer internship project.
Y. Hswen, Exploring discussions of health and risk
and public sentiment in Massachusetts during
COVID-19 pandemic mandate implementation: A
Twitter analysis, SSM Popul. Health 15 (2021)
100851.
[11]. Chakraborty, K.; Bhatia, S.; Bhattacharyya, S.;
Platos, J.; Bag, R.; Hassanien, A.E. Sentiment
Analysis of COVID-19 tweets by Deep Learning
Classifiers—A study to show how popularity is A. Praneeth Surya is currently studying B.Tech with
affecting accuracy in social media. Appl. Soft specification of Information Technology in NRI Institute of
Comput. 2020, 97, 106754. [Google Scholar] Technology. He done a summer internship project .He
[CrossRef] [PubMed] completed two NPTEL courses.
[12]. Shahsavari, S.; Holur, P.; Tangherlini, T.R.;
Roychowdhury, V. Conspiracy in the time of corona:
Automatic detection of COVID-19 conspiracy
theories in social media and the news. J. Comput.
Soc. Sci. 2020, 3, 279–317. [Google Scholar]
[CrossRef] [PubMed]
[13]. Havey, N.F. Partisan public health: How does
political ideology influence support for COVID-19
related misinformation? J. Comput. Soc. Sci. 2020, 3, D.Siva Teja is currently studying B.Tech with
319–342. [Google Scholar] [CrossRef] [PubMed] specification of Information Technology in NRI Institute of
Technology. He done a summer internship project. He
completed one NPTEL course.

IJISRT23APR1751 www.ijisrt.com 1445