Recognition of Fraud identities on Social Network using Convolutional Neural Network

  • Bharat Borkar, Manish Sharma

Abstract

Identity identification plays an important part in social network platforms, various platform are facing the problem of fake accounts  since many years in current years. Most of the authors has identified approach to find the fake profiles, but still  not able to identify the system which will be  find the complete solution for the  such issues. These fake identities are used by criminals for different malicious purposes, it  becomes necessary to identify them. The fake identities can be differentiated in two main types’ i.e. fake identities by bots and fake identities by humans. The purpose of this system is to  removes fake identities by bots during preprocessing and targets mainly on identification of fake identities by humans, as very little research has been made till now on the fake identities by humans. For classification we test for two different algorithms i.e. Random Forest (RF) and Convolution Neural Network (CNN). The identification is based on various features such as name, location, friends count, followers count and so on. Here, dataset used is that of Twitter. 

Published
2020-08-01
Section
Articles