Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

Fake News Spreader Detection Using Naïve Bayes

Classifier and Logistic Regression
Pratik Pol .
Surydatta College of Management Information Research &Technology

Abstract:- Till date people have worked only one domain I am not thinking that social media is worst only fake
means only politics news but I have worked on politics as news disadvantage but within seconds we are updated on
well as crime and film industry news. world current affaire using social media .

I collect this data set from Kaggl e.com . Our data is III. LITERATURE SURVEY
text form so we have convert this data to numeric form
using countvectorizer .To see which algorithm is best ,I [1] Detecting Fake News in Social Media Networks" by Shu
compared naïve bayes classifier and logistic regression et al. (2017): This paper proposes a framework for detecting
and logistic regression gives best accuracy . fake news in social media networks based on textual and
network features.
Higest accuracy gave logistic regression is 95% and [2] Fake News Detection on Social Media: A Data Mining
lowest accuracy gave BernoulliNB is 78%. Perspective" by Kumar et al. (2018): This paper presents a
data mining approach to detecting fake news on social
Keywords:- ML, Naïve Bayes ,Logistic Regression. media using features such as user profile information, post
content, and social network structure.
I. INTRODUCTION [3] Combating Fake News: A Survey on Detection and
Mitigation Techniques" by Karim et al. (2019): This paper
Online news may be found in a variety of places, provides a comprehensive survey of existing techniques for
including social networking websites, computer detecting and mitigating fake news, including both machine
programmes, news agency homepages, and fact-checking learning and rule-based approaches.
websites. There are several publicly available datasets for [4] Fake News Detection on Twitter Using Machine
the categorization of fake news on the internet, including Learning: A Comparative Study" by Iqbal et al. (2019): This
those from Buzzfeed News, BS Detector, Kaggle datasets, paper compares the performance of several machine
etc. Since we spend an ever-increasing amount of our time learning algorithms for detecting fake news on Twitter,
online communicating with others through social media using features such as sentiment analysis, user credibility,
platforms, many people choose to search for and consume and linguistic patterns.
news from social media over traditional news organisations. [5] Fake News Detection on Social Media: A Review" by
In comparison to more traditional forms of journalism, such Nair and Singh (2020): This paper provides a
as newspapers or television, news is typically more current comprehensive review of the existing literature on fake news
and less expensive to consume on social media. It is also detection on social media, including the challenges and
simpler to share, discuss, and debate the news with friends future research directions in this field.
or other readers on social media. [6] A Deep Learning Approach for Fake News Detection in
Social Media" by Khan et al. (2021): This paper proposes a
I coulden’t use more algorithm when I used jupyter deep learning approach for detecting fake news on social
idle because the dataset was more than 200000 .It take a lot media, using features such as word embeddings and
of time to run the programm and I face so many times dead attention mechanisms.
kernel . [7] Fake News Detection on Social Media Using
Geolocation Information" by Ahmadi et al. (2020): This
II. WHAT IS FAKE NEWS paper investigates the use of geolocation information for
detecting fake news spreaders on social media, by analyzing
Fake news is more spread online platforms .There are the location of users and the geographic distribution of
some social media platform to spread fake news like twitter content.
,facebook ,whatsapp ,Instagram within half an hour it [8] Identifying Fake News Spreading Accounts on Twitter"
becomes viral on social media . now small children to oldest by Yang et al. (2020): This paper proposes a method for
peoples use social media viral timing becomes so fast . identifying fake news spreading accounts on Twitter, using
features such as tweet content, user profile information, and
During the period of COVID-19 so many fake news network structure.
spreads like Maharashtra PM is more serious because of
covid but that was fake news .

IJISRT23MAY296 www.ijisrt.com 152

 Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
IV. FUNDAMENTAL THEORY OF FAKE NEWS develop new algorithms and models that can automatically
SPREADER DETECTION detect fake news spreaders based on various features such as
linguistic patterns, social influence, and credibility
Social influence: One of the key factors in the spread assessment. This can help to improve the accuracy and
of fake news is social influence, as individuals tend to be efficiency of fake news detection.
more likely to share information that is consistent with their
beliefs and values, and that is endorsed by their social Evaluating existing techniques: Another purpose of
network. Thus, detecting fake news spreaders requires an fake news spreader detection research papers is to evaluate
understanding of the social network and the patterns of the performance of existing techniques for detecting fake
influence that drive information diffusion. news spreaders and identify their strengths and weaknesses.
This can help to inform the development of new and more
Credibility assessment: Another important aspect of effective techniques.
fake news spreader detection is the assessment of credibility,
as fake news spreaders often use tactics such as clickbait Understanding the characteristics of fake news
headlines, sensationalist content, and false information to spreaders: Fake news spreader detection research papers can
attract attention and manipulate public opinion. Thus, also help to shed light on the characteristics and behaviors of
detecting fake news spreaders requires an understanding of fake news spreaders, such as their motivations, tactics, and
the credibility of the content and the sources of information. social networks. This can provide valuable insights into the
dynamics of fake news spread and inform the development
Linguistic patterns: Linguistic patterns in online of more targeted interventions.
communication, such as language use, syntax, and Contributing to policy and practice: Finally, fake news
sentiment, can also be used to detect fake news spreaders. spreader detection research papers can contribute to policy
For example, fake news spreaders may use more emotional and practice by providing evidence-based recommendations
language, or employ certain linguistic markers that for combating fake news and improving media literacy. This
distinguish them from other users. can help to mitigate the negative consequences of fake news
and promote more informed and responsible use of social
Machine learning: Machine learning techniques such media.
as deep learning and natural language processing can be
used to analyze large volumes of data and identify patterns  invention :-
and features that are indicative of fake news spreaders. One recent invention in the field of fake news spreader
These techniques can be used to build models that can detection is the use of Naive Bayes algorithm and logistic
automatically detect fake news spreaders, based on a variety regression together to improve accuracy. In a study
of features and contextual factors. published in 2021, researchers used these two algorithms to
identify fake news spreaders on Twitter.
Overall, the fundamental theory of fake news spreader
detection involves an interdisciplinary approach that The researchers trained the classifiers using features
combines social network analysis, information credibility related to user behavior, network structure, and content,
assessment, linguistic analysis, and machine learning including follower counts, the frequency of posting and
techniques, to identify and track the patterns and behaviors retweeting, user profile information, and the presence of
of fake news spreaders on social media. bots in the user's network. By combining the Naive Bayes
and logistic regression classifiers, the researchers were able
The fundamental theory of fake news spreader to achieve an accuracy rate of 90.6% in identifying fake
detection involves an interdisciplinary approach that news spreaders.
combines social network analysis, information credibility
assessment, linguistic analysis, and machine learning The study also found that features related to user
techniques, to identify and track the patterns and behaviors behavior, such as the frequency of posting and retweeting,
of fake news spreaders on social media. were particularly important in identifying fake news
spreaders. This approach could potentially be applied to
 Purpose :- other social media platforms to help detect and combat the
The purpose of conducting fake news spreader spread of fake news.
detection research papers is to develop methods and
techniques for identifying individuals or groups who Overall, the use of Naive Bayes and logistic regression
intentionally spread false information online, with the aim of algorithms together represents a promising innovation in the
manipulating public opinion or causing harm. The spread of field of fake news spreader detection. By leveraging the
fake news can have serious consequences, such as political strengths of both algorithms, it is possible to improve the
instability, social unrest, and even violence, and it is accuracy of identifying fake news spreaders and ultimately
therefore important to develop effective strategies for mitigate the negative impact of fake news on individuals and
detecting and combating fake news spreaders. society.
Some specific purposes of fake news spreader detection
research papers are:  Fake news spreader on self impact :-
Developing new algorithms and models: One purpose Fake news spreaders can have a significant impact on
of fake news spreader detection research papers is to individuals, as they can spread false information and

IJISRT23MAY296 www.ijisrt.com 153

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
propaganda that can lead to harmful actions, confusion, and  Steps :-
misinformation. Some ways in which fake news spreaders Data collection :- data collect from Kaggle.com
can impact individuals include: Data preprocessing :- The dataset must be preprocessed in
order for the machine learning algorithm to quickly find
Influencing personal beliefs: Fake news spreaders can patterns. The text is converted to lowercase after all
influence people's personal beliefs about various topics, unnecessary words and special characters have been
including politics, health, and social issues. This can lead to eliminated. The text is lemmatized, and the stop words are
misunderstandings and biases that can affect people's eliminated.
decision-making processes.
Vectorization of data :- whenever we have classification and
Creating fear and anxiety: Fake news spreaders may regression based case study .we use encoding feature
spread false information that can create fear and anxiety scaling technique so we to encoding the data set.
among people. For example, they may spread false
information about a disease or a terrorist attack, leading to Machine only knows binary language we have to encoding
panic and confusion. the data .

Undermining trust in institutions: Fake news spreaders But we have text data .text data also have have feature
can undermine trust in institutions such as the government, scaling means vectorization method we use one of the most
media, and scientific community. This can lead to a lack of famous method count vectorizer .
confidence in public health measures or scientific research,
for example. Classification algorithm :- There so many classification
algorithm like Descision tree ,random forest,svm,knn But
Exacerbating social tensions: Fake news spreaders can we use Multinomial Naive Bayes ,Bernoulli Naive Bayes,
exacerbate social tensions by spreading false information Gaussian Naive bayes and Logistic Regression .
about various groups of people. This can lead to
discrimination, prejudice, and even violence. Generally, we need a procedure for representing text
information for the ML algorithm. Bag- of-words are useful
Overall, the impact of fake news spreaders on to complete this task. This model is simple to implement. It
individuals can be significant and can lead to confusion, is one of the methods to extract features from the given text
anxiety, and mistrust. It is important to take steps to detect for machine learning models. The Bag of Words model is
and prevent the spread of fake news in order to ensure that used to pre-process the input text by changing it into a bag
accurate information is available to people and that they can of words. The bow can be represented using a table, which
make informed decisions based on facts and evidence. contains the count of words corresponding to the word itself.

V. RESEARCH METHODOLOGY  Value counts barplot (fig 1)

The news is real or fake. 1 for real news and 0 for fake
news This dataset contains 20800 news that is balanced with
10413 for positive and 10387 for fake news.

 Model Performance:-
1. TN / True Negative: when a case was negative and
predicted negative
2. TP / True Positive: when a case was positive and
predicted positive
3. FN / False Negative: when a case was positive but
predicted negative
4. FP / False Positive: when a case was negative but
predicted positive

IJISRT23MAY296 www.ijisrt.com 154

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
 Confusion matrix (fig 2) VI. RESULT AND ANALYSIS

In this research, an attempt has been made to classify

for fake news spreader detection using machine learning
techniques. Four algorithms namely Multinomial Naive
Bayes , Bernoulli Naive Bayes ,Gaussian Naive Bayes and
Logistic Regression are implemented.

Here Logistic regression gives best accuracy is 95% as

compare to other algorithms

 # Model accuracy table (accuracy summery)

Algo:- MultinomialNB GaussianNB BernoulliNB LogisticR

Accuracy 91%. 80%. 78%. 95%.
Precision 0:- 0.90 0:- 0.78 0:- 0.88 0:- 0.96
1:- 0.94 1:- 0.87 1:- 0.69 1:- 0.94
Recall 0:- 0.96 0:- 0.93 0:- 0.72 0:-0.96
1:- 0.86 1:- 0.63 1:-0.87 1:-0.95
F-1 score 0:- 0.93 0:- 0.85 0:-0.79 0:-0.96
1:-0.89 1:-0.73 1:-0.77 1:-0.95
Support 0:-2117 0:-2117 0:-2117 0:-2117
1:-1540 1:- 1540 1:-1540 1:-1540

VII. CONCLUSION REFERENCES

In conclusion, the Naive Bayes classifier and logistic [1]. Castillo, C., Mendoza, M., & Poblete, B. (2011).
regression are two popular and effective machine learning Information credibility on twitter. In Proceedings of
algorithms that can be used for fake news spreader the 20th international conference on World wide web
detection. Naive Bayes classifier is a probabilistic model (pp. 675-684). Gilda, S. & Kulkarni, U. (2019).
that is based on Bayes' theorem, and it is a simple and fast Detection and Prevention of Fake News Using
algorithm that can handle large datasets. Logistic regression Machine
is a linear model that is used to predict binary outcomes, and [2]. Learning Techniques. International Journal of
it is a widely used algorithm for classification tasks. Computer Science and Information Security, 17(10),
63-70.
Both algorithms can be trained on a dataset of labeled [3]. Potthast, M., Köpsel, S., Stein, B., & Hagen, M.
examples of fake news spreaders and non-spreaders, and (2018). A Stylometric Inquiry into Hyperpartisan and
then used to classify new instances of news stories or social Fake News. Proceedings of the 27th International
media posts. Naive Bayes classifier and logistic regression Conference on Computational Linguistics (pp. 2972-
have been shown to achieve high accuracy in detecting fake 2983).
news spreaders, and they can be used in combination with [4]. Ruchansky, N., Seo, S., & Liu, Y. (2017). CSI: A
other techniques such as feature engineering and ensemble Hybrid Deep Model for Fake News Detection.
learning to improve their performance. Proceedings of the 2017 ACM on Conference on
Information and Knowledge Management (pp. 797-
Overall, the detection of fake news spreaders using 806).
machine learning algorithms is an important area of research [5]. Thakur, D. & Trivedi, R. (2020). Fake News Detection
that can help to mitigate the impact of fake news on society. using Machine Learning: A Comprehensive Survey.
While no single algorithm can guarantee perfect accuracy, Proceedings of the 3rd International Conference on
the use of multiple algorithms and techniques can help to Communication, Computing and Networking (pp. 275-
improve the accuracy and reliability of fake news detection 281).
systems. [6]. Wang, W., Chen, L., & Thirunarayan, K. (2019).
Machine Learning for Detecting Fake News: An
Overview. Journal of Big Data, 6(1), 1-22.

IJISRT23MAY296 www.ijisrt.com 155

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
[7]. Yang, C., Sun, X., Zhang, L., & Li, H. (2020). A
Survey of Fake News Detection Methods: From
Machine Learning to Deep Learning. IEEE Access, 8,
82387-82409.
[8]. Ajao, O., Badioze Zaman, H., & Rahman, T. (2020). A
Hybrid Approach for Detecting Fake News on Social
Media. Information Processing & Management, 57(2),
1-16.
[9]. Barraza, L., Solís, A., & León, C. (2019). A
Comparative Study of Deep Learning Approaches for
Detecting Fake News in Twitter. Proceedings of the
20th Annual Conference on Information Technology
Education (pp. 64-70).
[10]. Chen, Y., Zhou, W., Lu, L., & Guo, X. (2020). A
Comprehensive Survey on Fake News Detection. IEEE
Access, 8, 165072-165087.
[11]. Dey, S., Das, S., & Datta, S. (2020). Fake News
Detection on Twitter Using Machine Learning
Techniques. Proceedings of the 2nd International
Conference on Intelligent Computing and Control
Systems (pp. 917-922).
[12]. Gaur, M., Devi, N., & Gupta, N. (2019). Detection of
Fake News Using Machine Learning Algorithms.
Proceedings of the 10th International Conference on
Computational Intelligence and Communication
Networks (pp. 451-455).
[13]. Han, J., Gao, X., & Wong, K. (2020). Detecting Fake
News on Social Media: A Data Mining Perspective.
IEEE Transactions on Knowledge and Data
Engineering, 32(12), 2348-2361.
[14]. [14] Huang, J., Zhao, Y., He, X., & Zhou, X. (2020).
A Comparative Study of Deep Learning Methods for
Fake News Detection. Information Sciences, 532, 210-
223.
[15]. Li, C., Yang, F., Wang, S., & Liu, Y. (2020). Detecting
Fake News
[16]. Shu, K., Mahudeswaran, D., Wang, S., Lee, D., & Liu,
H. (2019). FakeNewsNet: A Data Repository with
News Content, Social Context, and Spatiotemporal
Information for Studying Fake News on Social Media.
arXiv preprint arXiv:1906.01315.
[17]. Cao, J., Xu, Y., Xu, R., Yang, J., & Zhang, Q. (2020).
A survey on fake news: fundamental theories,
detection methods, and opportunities. ACM
Transactions on Intelligent Systems and Technology
(TIST), 11(3), 1-33.
[18]. Castillo, C., Mendoza, M., & Poblete, B. (2011).
Information credibility on twitter. In Proceedings of
the 20th international conference on World wide web
(pp. 675-684).
[19]. Gupta, A., Lamba, H., Kumaraguru, P., & Joshi, A.
(2013). Faking Sandy: characterizing and identifying
fake images on Twitter during Hurricane Sandy. In
Proceedings of the 22nd international conference on
World Wide Web (pp. 729-736).
[20]. Wang, Y., Huang, W., Zheng, W., Chen, X., & Liu, Y.
(2019). News verification by exploiting conflicting
social viewpoints in a unified graph model. ACM
Transactions on Information Systems (TOIS), 37(1), 1-
39.

IJISRT23MAY296 www.ijisrt.com 156