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

ISSN No:-2456-2165

E-Mail Spam Detection by using NLP and Naïve

Bayes Classification Through Machine Learning
Mr. A. Ravi Kiran1 T. Sai Sowjanya2
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. V. CH. Sai Pavan 3 I. Naveena4

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:- Internet has become to be an integral part of systems. Its miles envisioned that junk mail fee
lifestyles. With multiplied use of internet, numbers of e- Organizations on the order of one hundred billion dollars in
mail customers are growing day by day. This growing 2007.
use of e-mail has created troubles induced. Through
unsolicited bulk email messages normally called spam. On this challenge, we use textual content mining to
Electronic mail has now come to be one of the perform automated junk mail filtering to use emails
satisfactory methods for commercials due to which junk effectively. Most spam emails divert people’s attention
mail emails are generated. Unsolicited mail emails are away from genuine and important emails and direct them
the emails that the receiver does not preference to towards detrimental situations.
gather. A massive quantity of equal messages is sent to
numerous recipients of e-mail. Direct mail usually arises We strive to pick out Styles using statistics-mining
as a result of giving out our email address on an type algorithms to allow us classify the emails As HAM or
unauthorized or unscrupulous internet website. There SPAM. [Fig:1]
are a few of the consequences of junk mail. Fills our
Inbox with type of ridiculous emails, that will reduce our To solve this trouble, numerous unsolicited mail
internet speed. Steals beneficial records like our info on detection techniques are used now. The Most commonplace
you contact list. Alters your seek consequences on any approach for junk mail detection is the usage of Naïve
laptop software. Junk mail is a huge waste of every Bayesian method and function sets That verify the presence
body’s time and can quickly turn out to be very of unsolicited mail key phrases in the incoming mil.
frustrating if you get hold of big quantities of it. Figuring
out these spammers and the junk mail content is an
onerous challenge. Despite the fact that full-size variety
of studies were executed, but to this point, the
Techniques set forth nevertheless scarcely distinguish
spam surveys, and none of them show the Benefits of
each eliminated detail compose. Despite increasing
network verbal exchange and Losing lot of reminiscence
space, spam messages also are used for some attacks.

Keywords:- Spam Detection, Python, Machine Learning,

Stop Words, Tokenization, Potter Stemmer, Count
Vectorizer, Label Encoder, Naïve Bayes Algorithm, NumPy,
Pandas.

I. INTRODUCTION

In these days’ s globalized international, e-mail is a

Fig 1 ML model for Spam Detection
number one source of conversation. This communication
can vary from personal, commercial enterprise, corporate to
government. With the rapid growth in email utilization,
there has additionally been growth in the unsolicited mail
emails. Spam emails, additionally referred to as junk e-mail
includes nearly equal messages dispatched to several
recipients via electronic mail. Apart from being stressful,
spam emails can additionally pose a safety chance to our

IJISRT23MAY321 www.ijisrt.com 147

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
II. TECHNOLOGIES USED  Word Tokenize.
 NumPy.
 Pandas.
 Porter Stemmer.
 Stop words.
 Label Encoder.
 Count Vectorizer.

 Naive Bayes:
It is a supervised learning-based algorithm, this
algorithm is mostly used for classification areas like text
classification. It is the simplest and effective algorithm that
allows the ML models to make faster predictions.

The naïve bayes is a Probabilistic classifier i.e., it

predicts the output based on the probability of an object in
the text or input data. [Fig:3]

Fig 2 Technologies used

 Machine Learning:
Machine Learning is a way of training a model to be
able to understand the pattern in the output with respect to
the output. In one word it is an advanced algorithm that can
process the data and can generate a AI program that can
capable to run independently in individual systems with
different inputs. Now-a-day machine learning being used in
everywhere.

Fig 4 Naïve Bayes

 Natural Language Processing:

Natural language processing (NLP) refers to the
branch of Computer science and more especially, the
department of Artificial intelligence or AI concerned with
giving computers the ability to understand textual content
and spoken words in a good deal the same manner humans
can understand. [Fig:4]

NLP combines computational linguistics rule-based

modeling of human language with statistical, gadget
studying, and deep gaining knowledge of models.
Collectively, those technologies allow computers to method
human language in the form of textual content or voice
statistics and to ‘apprehend’ its complete which means,
Fig 3 Machine Learning entire with the speaker or creator’s intent and sentiment.
 Python:
Python is a totally popular general-cause interpreted,
interactive, object-orientated, and high-level programming
language. Python is dynamically-typed and programming
language. It created by means of Guido van Rossum in the
course of 1985- 1990.
Python includes various machine mastering libraries that
could cope with massive files and can get entry to every
phrase in it. [Fig: 2]

In our project we have used the python libraries like

NUMPY and PANDAS that has several sub-libraries that
are as follows. Fig 5 NLP

IJISRT23MAY321 www.ijisrt.com 148

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
 NumPy and Pandas: IV. EXISTING SYSTEM
NumPy and Pandas are the libraries of Python, these
are used for the purpose of data analysis operations in The methods are currently used by the most anti-spam
Python. NumPy is used while working with the numerical software are static, mean that it is fairly easy to avoid by
data as it makes easy to handle mathematical functions like them to tweaking the message a little. To effectively battle
matrices and multi-dimensional arrays. Pandas is an Open- the spam, an adaptive new technique is needed. Any small
source Python package, it is mostly used for the data modification in the data can easily manipulate the spam
analysis and data science and ML operations. filter. This method must be familiar with spammer’s tactics
as they change over time.
Another main advantage of NumPy and Pandas are,
these two Libraries have high Processing Speed. Sometimes the important mails are also getting into
spam folder, this is because of, these spam filters more
likely to consider the repeated sender as the spam, which
may contain the important data. [9].

 Disadvantages:

• Time consuming.
• Memory wastage.
• No standard Classifier.
• Filtration done based on sender.
• Less accuracy.
• Small modifications in the incoming mail can easily
manipulate the filter.
• Sometimes Ham mails are also sent to spam folder.

V. PROPOSED SYSTEM

In this we are proposing a system which detects the

spam messages using NLP (Natural Language Processing)
Fig 6 NumPy and Pandas
Our system mainly deals with the format of the spam mail
data and will try to analyze the information by using ML
III. SOFTWARE REQUIREMENTS
(machine learning) algorithms, we will make use of Natural
SPECIFICATION
language processing technique train a model and perform
spam detection. here we follow the various ways to process
 Python 3.0 or later or Jupiter Notebook.
the data by using Tokenization, Stemming, StopWords. etc.
 Windows XP, 7, 8, or 10 operating system. In our project we will train the ML model the
 NumPy, pandas, sklearn, and matplot; and an understand the format of the spam mails and the ham mails
 Internet browser (google chrome or Firefox) (not a spam mail) the mostly used words in spam and ham
 8GB ram and minimum ROM with Intel i3 or higher. are analyzed, the accuracy could be increased by given new
spam mails to ML algorithm. It is utmost necessary to stop
all the unwanted messages as they contain viruses which
harms the computer. It is reliable means it can be accessed
in multiple systems.

This helps the user to avoid the spam messages by

predicting the spam mail before it entering into the Inbox.
The ML and NLP model can produce accurate results when
compared to previous models i.e., LR model (Logistic
Regression). We are going to develop this model by using
all the above concepts.

 Advantages of Proposed System:

• High Accuracy.
• Block mail from known spam sources.
• It is effective and easy to implement.
• The presence of single Token should not cause the e-
Fig 7 Software Requirement Specifications mail to be classified as spam.
• Low rate of false positives.

IJISRT23MAY321 www.ijisrt.com 149

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
VI. SYSTEM ARCHITECTURE VII. CONCLUSION

Spam mails are a serious concern a major displeasure

for many Internet users. The mode proposed as a solution in
this paper is highly beneficial because it introduces a
threshold counter that helps overcome congestion on the
web server and also maintain the spam filter capability, but
at the same time, it also requires overhead storage space for
the databases.

Since NLP is a relatively underdeveloped area for

research, further enhancements can be made in the field of
spam detection for online security using Natural Language
Processing in the future.

REFERENCES

[1]. Qaroush, A., Khater, 1. M., &Washaha, M. (2012).

Identifying spam e-mail based-on statistical header
features and sender behavior. In Proceedings of the
CUBE international information Technology
Conference, pp. 771-778.J. Wong, “Almost all the
traffic to fake news sites is from facebook, new data
show,” 2016. View at: Google Scholar
[2]. M. Basavaraj, R. Prabhakar, "A Novel Method of
Spam Mail Detection using Text Based Clustering
Approach", International Journal of Computer
Applications (0975–8887), vol. 5, no. 4, August
Fig 7 System Architecture 2010..
[3]. Comprative study of classification Algorithms for
 Class Diagram: spam Email dection Aakankasha sharaff, Naresh
Kumar Nagwani & Abhishek Dhadse. View at:
Google Scholar
[4]. C. Baziotis, N. Pelekis, and C. Doulkeridis,
“Datastories at semeval-2017 task Deep lstm with
attention for message-level and topic-based sentiment
analysis”, in Proceedings of the 11th international
workshop on semantic evaluation (SemEval-2017),
2017, pp. 747–754.
[5]. G. Egozi and R. Verma, “Phishing email detection
using robust nlp techniques”, in 2018 IEEE
International Conference on Data Mining Workshops
(ICDMW), IEEE, 2018, pp. 7–12.. View at: Google
Scholar
[6]. Karthick veerakumar, Spam filter, 2017. [Online].
Available:
https://www.kaggle.com/karthickveerakumar/spam-
filter. View at:Google Scholar
[7]. Spam detection Detection Using Machine Learning
Ensemble Methods (hindawi.com)
[8]. Spam Detection Using Machine Learning Algorithms
(researchgate.net)
[9]. E. Blanzieri and A. Bryl, “A survey of learning-based
techniques of email spam filtering,” Artificial
Fig 8 Class Diagram Intelligence Review, vol. 29, no. 1, pp. 63–92,
2008.View at: Publisher Site | Google Scholar
 Future Scope:
In future, we would like to update our project with the
advanced algorithms by using Neural networks, that are
having the capability to update themselves for the new
modified spam mails. So that the space complexity can be
reduced to minimum.

IJISRT23MAY321 www.ijisrt.com 150

Volume 8, Issue 5, May – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
[10]. N. Udayakumar, S. Anandaselvi, and T.
Subbulakshmi, “Dynamic malware analysis using
machine learning algorithm,” in Proceedings of the
2017 International Conference on Intelligent
Sustainable Systems (ICISS), IEEE, Palladam, India,
December 2017.View at: Google Scholar
[11]. S. O. Olatunji, “Extreme Learning machines and
Support Vector Machines models for email spam
detection,” in Proceedings of the 2017 IEEE 30th I. Naveena is currently studying B.Tech with
Canadian Conference on Electrical and Computer specification of Information Technology in NRI Institute of
Engineering (CCECE), IEEE, Windsor, Canada, Technology. She done summer internship on Spam
April 2017.View at: Google Scholar Detection.
[12]. M. H. Arif, J. Li, M. Iqbal, and K. Liu, “Sentiment
analysis and spam detection in short informal text
using learning classifier systems,” Soft Computing,
vol. 22, no. 21, pp. 7281–7291, 2018.

BIOGRAPHIES

Mr. Aala Ravikiran is presently working as Assistant

Professor in the department of Information Technology at
NRI Institute of Technology, Vijayawada. He received his
M. Tech degree from Jawaharlal Nehru Technological
University, Kakinada (JNTUK). He has published over 2
research paper in international journals. He has more than 6
years of experience in Teaching.

T. Sai Sowjanya is currently studying B.Tech with

specification of Information Technology in NRI Institute of
Technology. She done summer internship on Spam
Detection.

A. V. CH. Sai pavan is currently studying B.Tech with

specification of Information Technology in NRI Institute of
Technology. He done summer internship on Spam
Detection.

IJISRT23MAY321 www.ijisrt.com 151