.

ISSN 2063-5346
For urgent queries please contact : +918130348310

DETECTION OF SOCIAL NETWORK SPAM BASED ON MACHINE LEARNING WITH NAIVE BAYES ALGORITHM

Main Article Content

Dr. K. Anuradha, Dr. T. Guhan, Dr. N. Revathy, Dr. K. Jegadeeswaran
» doi: 10.31838/ecb/2023.12.s3.265

Abstract

Social media is an internet platform that allows users to effortlessly build social connections with other users (Facebook, Email, LinkedIn). People today share their personal information, habits, career interests, and hobbies with their peers on social networking platforms. With online social networks, different information is spreading, both good and bad. Sharing information online is getting more commonplace every day. The naive Bayes method is utilised in this study to identify false information, such as internet rumours, as well as to solve distinguishing and predicting issues. Moreover, the Naive Bayes method is utilised for collaborative filtering, hybrid recommender systems, spam filtering, and text categorization. The premise behind many social networks is that a user's online information represents who they really are. Members of these networks who fill in their name fields with made-up names, corporate names, phone numbers, or just random characters are breaking the terms of service, tarnishing search results, and lowering the site's value for legitimate users. Identifying and banning these accounts based on their spammy names might enhance actual users' experiences on the site and stop more abusive behaviour. This project's primary goal is to identify and categorise email communications into spam and junk using various machine learning approaches. The NLP needs to be used (Natural Language Processing). This study aims to understand how different machine learning algorithms can categorise email spam with ham. We must now create machine learning-based techniques for identifying email spammers. Any emails that include unsolicited content and show up in a user's email box are referred to as spam. Spam is frequently responsible for network bottlenecks, blocking, and even harm to the electronic messaging system. We must integrate several machine learning techniques int our workflow, including support vector machines and naive bayes.

Article Details