Dena: Using social network goals and machine learning to detect fake accounts and improve social media security

Today, with the pervasiveness of social networks, the security of this environment is considered one of the most important issues. One of the security challenges is the creation of fake accounts that harass social media users. The owners of these fake accounts pursue goals such as creating likes and followers or distributing misinformation for political, cultural and economic purposes. In this study, with the aim of improving security in social networks and improving the security of cyberspace, a method for investigating and detecting fake accounts will be presented. The proposed method called "Dena" will use the goals of the social network on the one hand and the combined algorithm method with the decision tree, the nearest neighbor and Bayes on the other hand.. The results of the proposed method with a hybrid algorithm show an accuracy of 95.34%. Stability and lack of overfit are other features of the proposed method that have been proven in the results section. The results of this study can be used to provide solutions to prevent the creation of fake accounts and increase its security and lead to the recognition and use of new data mining techniques in social networks and be used in the field of data analysis and data mining in social networks.