Twitter Fake Accounts Identification Using Support Vector Machine

Abstract

In this work the impact of the advanced social media platforms is clearly described. The techniques used to identify the twitter accounts were first discussed to conclude a better state of the art mechanism would be developed. After analyzing the progressive approaches already being in use to overcome the fake social media accounts and realizing the demand of social media platforms as the data is increasing exponentially. As the exponential increase in the user of twitter similarly the fake accounts also increased. There are different types of analysis are conducted on twitter data sets, content-based, URL based, fake user detection. After analyzing all the methods it uses a support vector machine on the cresci-2017 data set of twitter and used twenty features to conclude results, containing improved accuracy and faster results. Fake accounts are identified by analyzing the friend counts and the follower's count of the accounts. Similarly, it is also concluded that most fake accounts have less likely to fill their location data, less likely to add status and extend the profile. The results were observed as improved and fast using a support vector machine