click here to download the abstract project:
click here to download the base paper
ABSTRACT:
Email is a medium of information that is still frequently used by people today. At the moment email still has an endless problem that is spam email. Spam email is an email that can pollute, damage or disturb the recipient. In this study, we show the performance and accuracy of Multinomial Naïve Bayes (MNNB) and Complete Gini-Index Text (GIT) for use in spam email filtering. In this study, we used 6 cross-validations as testers for the built classification machines. We found that the average yield can exceed Multinomial Naïve Bayes without using feature selection which only uses 80000 features with a difference of 0.39%. Feature selection also increases speed during classification and can reduce features that are less relevant to the category to be classified.
INTRODUCTION:
The Internet has become a common thing in our lives.The same message sends multiple times which affects the organization financially and also irritates the receiving user. In this project, a Spam Mail Detection system is proposed will classify the given email as spam or ham email. Spam filtering mainly focuses on the content of the message. The classification algorithm classifies the given email based on the content. Feature extraction and selection plays a vital role in the classification. In spam mail detection, email data is collected through the dataset. To obtain the accurate results, data needs to be pre-processed by removing stop words and word tokenization. Pre-processing of data is done by using TF-IDF Vectorizer module. SVM algorithm is used to detect the given email is spam or harm. In recent times, unwanted industrial bulk emails known as spam has become an enormous drawback on the net. The person causing the spam messages is noted because the sender. Such an individual gathers email addresses from completely different websites, chatrooms, and viruses. Spam prevents the user from creating full and sensible use of your time, storage capability and network information measure. the massive volume of spam mails flowing through the pc networks have damaging effects on the memory house of email servers, communication information measure, central processing unit power and user time. The menace of spam email is on the rise on yearly basis and is to blame for over seventy-seven of the entire international email traffic. Users United Nations agency receive spam emails that they failed to request realize it terribly irritating. it’s conjointly resulted to much loss to several users United Nations agency have fallen victim of web scams and different dishonest practices of spammers United Nations agency send emails pretence to be from honorable firms with the intention to influence people to disclose sensitive personal info like passwords, Bank Verification variety (BVN) and mastercard numbers.
