Fake profile detection in social media using image processing and machine learning

Abstract

Almost everybody has a social media presence in today’s technologically advanced world. As a result, making fake accounts is very easy. The term ”fake profile” refers to a person who may pretend to be someone else. These accounts are mostly used to impersonate others and defame them. Furthermore, a fake account can be used for various reasons, including igniting political feuds, spreading misleading facts, and disseminating news about current sensitive topics. Since fake profiles pose such a serious threat to everyone, a model was proposed that might aid in the reduction of fake profiles. It can assist with identifying accounts that could be accused of being fraudulent, such as those without a profile photo. To ensure that each user has a unique profile, machine learning and image recognition was used in our model. Our model attempted to discourage users from creating accounts using the photo or knowledge of another person. To do this, One Time Password (OTP) was implemented so that fake users can not get the chance to create an account by using another person’s name. Fake accounts needed to identify by using deep learning from a real dataset of people’s answers. To detect the false results, the k-means algorithm was implemented on our dataset. When the k-means clustering algorithm was used on the dataset, it was discovered that our detection accuracy was 75.30 percent.