International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 03 | Mar 2025
p-ISSN: 2395-0072
www.irjet.net
Detecting Spam Email with Machine Learning and Writing Optimized Email Siddharth More1, Rajat Pandit2, Yash Salunke3, Rohit Jawale4, Prof. Dr.Radhika Nanda5 1,2,3,4 B.E. Students Department of Computer Engineering
5 HOD, Department of Computer Engineering, Bharat College of Engineering, Opp. Gajanan Maharaj Temple,
Kanhor Road, Badlapur (West), Thane, Maharashtra - 421503 ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract – In today's digital era, emails have become a
evade detection further. To counter this, Optical Character Recognition (OCR) technology was developed to extract text from images, allowing traditional text-based filtering techniques, such as Naïve Bayes and other classification methods, to be applied. Despite these advancements, spammers continually evolve their strategies to bypass security measures, making spam detection an ongoing challenge.
fundamental mode of communication for both personal and professional purposes. However, with the growing number of email users, the volume of spam emails has also increased. Spam, often referred to as junk mail, consists of unwanted messages sent in bulk to multiple recipients, typically for commercial gain. These emails may include phishing attempts, image-based spam, malware, lottery scams, advertisements, and other unsolicited content. Spam not only clutters inboxes with irrelevant messages but also poses security risks. It can slow down internet performance and enable cybercriminals to extract sensitive information, such as personal details, professional contacts, and financial data. As a result, distinguishing spam from legitimate (ham) emails is crucial. To combat this issue, spam filters are employed to identify and block unwanted, unsolicited, and potentially harmful emails before they reach the inbox. Various machine learning and deep learning techniques are used for spam detection, with some of the most effective being convolutional neural networks (CNN), support vector machines (SVM), and naïve Bayes (NB). Spam Email, phishing scams, unwanted communications, machine learning, convolution neural network CNN, support vector machine SVM, naïve bayes NB.
Fig. 1: Spam Filter
1. INTRODUCTION 1.1 Purpose
The Internet has become an integral part of modern society, enabling seamless communication and global connectivity at any time and from any location. One of the most widely used internet-based communication tools is email (electronic mail), which is utilized by students, professionals, business people, and government officials. Since sending emails is typically free, spammers exploit this feature to distribute large volumes of unwanted messages.
This discussion focuses on how machine learning can enhance spam email detection while improving email optimization. As email usage continues to grow, spam emails have become a major concern, leading to security threats and overcrowded inboxes. Machine learning techniques, including convolutional neural networks (CNN), support vector machines (SVM), and naïve Bayes (NB), play a crucial role in accurately identifying and filtering spam messages. Furthermore, crafting well-structured and optimized emails helps prevent legitimate messages from being mistakenly marked as spam.
Initially, most spam emails contained plain text, however, as text-based spam filters improved by analyzing email headers, body content, and other features, spammers adapted by embedding text within images—a technique known as image spam. They began incorporating backgrounds, noise, and other distortions into images to
© 2025, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 646