FAKE ACCOUNT DETECTION FROM URL FEATURES USING MACHINE LEARNING AND DEEP LEARNING

Abstract

Currently, social media has a significant impact on the lives of individuals. On a daily basis, the
bulk of individuals are dedicating their time to social media sites. The number of accounts on these social
networking sites has been steadily expanding on a daily basis, and a significant portion of users are engaging
with others regardless of their time and place. These social media platforms offer advantages and disadvantages,
and they also pose security risks to our personal information. In order to identify the individuals responsible for
making threats on these social networking sites, it is necessary to categorize the accounts into authentic and
fraudulent ones. Historically, many categorization approaches have been used to identify counterfeit accounts on
social media. However, it is essential to enhance the precision of spotting counterfeit accounts on these
platforms. Our work utilizes Machine Learning technologies, namely Deep Learning and Natural Language
Processing (NLP), to enhance the accuracy of identifying bogus accounts based on URL attributes. We choose
to use the Random Forest tree classifying technique.