ENHANCING SOCIAL NETWORK SECURITY: SPAMMER DETECTION AND FAKE USER IDENTIFICATION

Abstract

This project aims to delineate a methodology for detecting spam tweets and fraudulent user accounts on the social
network Twitter. We are utilizing a Twitter dataset and employing four distinct techniques: Fake Content Detection,
Spam URL Detection, Spam Trending Topic Identification, and Fake User Identification. By employing the
aforementioned four techniques, we can ascertain whether a tweet is normal or spam. Subsequently, we will utilize the
Random Forest data mining algorithm to train the dataset for the classification of spam versus non-spam tweets, as well
as fake versus non-fake accounts. We employ various data mining techniques to classify tweets as spam or non-spam,
specifically utilizing the Random Forest classifier in this instance. There is a demand to address and regulate individuals
who utilize online social networks solely for advertising, thereby spamming others' accounts. The recent identification
of spam on social networking platforms has garnered the interest of researchers. Spam detection poses a significant
challenge in safeguarding the security of social networks. Recognizing spam on OSN sites is crucial to protect users
from diverse malicious attacks and to safeguard their security and privacy.