International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 09 Issue: 06 | June 2022
p-ISSN: 2395-0072
www.irjet.net
E-Mail Spam Detection Using Supportive Vector Machine Muskan Dhirwani1, Kartikey Tiwari2, Mandeep Singh Narula3
3
1,2 Students, Dept. of ECE, Jaypee Institute of Information Technology, Noida, India Assistant Professor, Dept. of ECE, Jaypee Institute of Information Technology, Noida, India
------------------------------------------------------------------***----------------------------------------------------------------------Abstract -
With the rapid pace of growth of the internet it has been the most tedious job to classify the messages into ham and spam. Spam is defined as an unwanted message sent to somebody via email and messages. Many messages or emails often contain viruses, phishing material in order to break the privacy of the person by looking into its various confidential information. Many viruses try to get into the computer by this process of spam messages, so we have tried to build our project using the SVM process so that we could increase the spam detection feature of the model using neural networks in order to reduce the error. The features involved include regression, naive bayes and several other regression models.
Keywords: CNN, unwanted messages, viruses, filtration. 1. INTRODUCTION Messages in this present phase has changed its recognition, gone are the days when people used to message their friends with old simple monotonous simple text. But today after so much advancement with the time email has come into the playground. After the advancement of the internet tools like Artificial Intelligence have more power to play and influence the masses. This power either could be positive or it could be negative. In recent times email has been used for advertisement click bait videos and for several other promotional activities. In this project we have tried to build a model using Machine Learning which could segregate the spam mails which are sent to us. An email consists of three parts. The 1st consists of the source address, beneficiary location and topic. The 2nd comprises the main part or the body. The 3rd part might contain reports, pictures, sound or video documents. With the acceleration of the web and huge use of email there's an ascent inside the pace of spam messages. It consists of undesirable, excluded or useless messages. Programmers and other illegal persons try to break the sovereignty of the country using various methods to get into the inner details of the person like giving them clickbait messages, videos lure them with the spam messages. But recently engineers have © 2022, IRJET
|
Impact Factor value: 7.529
developed methods and procedures to find these types of emails. The advantage of these types of secure models are many like they try to avoid the spam and clickbait videos which have been sent to the person in order to carry out the illegal activities like phishing and spying into the system. This type of model allows the model to make the stock of the harmful words which allows the user to know about the lethalness of the virus and other spams. The offensive emails are sent to the people with their standard activity on the internet. By researching these principles one can work on the basic working of the model so that and so forth such that they can work on the basics of the spam. The model tries to reduce the spam finding value by several percent. In the learningbased work we have tried to implement the model using neural networks which would try to increase the spam detection ability of the model by various percentages. These models have tried to incorporate spam subjectivity, language identification and several other keywords which would try to increase the efficiency of the model. Thus, in this paper we have tried to build a model which could allow us to do spam detection using models like linear regression, naïve bayes, SVM.
2. BACKGROUND The paper revolves around the utility of the Support Vector Machine tool in Email detection. These days emails are often used to send important and crucial messages and information. But with the pace and growth of the internet comes the danger of cybercrime which involves techniques like phishing and bugs etc. These techniques are employed by the technique of spam mails. Just by sending spam mails and making users tempted to attempt it, this creates the problem for the user as it steals the important data, images, records and contacts from the user to the other person. This paper plans to use the analytics and model of SVM to carry out spam detection. With the rapid growth and development of the internet and the importance of mobility, emails have now become the tool of daily use. In the earlier days emails were mostly messages of plain text but as of now it comes with the hyperlinks, videos, sound and the other tempting tricks. An email message is broadly classified into three parts. The first part consists of the source address, sender address and subject. The second includes the message and the information which contains the most important part of an email. The third could contain reports, pictures, sound or video archives. With the speed |
ISO 9001:2008 Certified Journal
|
Page 370