Detection of Fake Instagram Account using XGBoost and Random Search

Instagram has been one of the most used social networks in the last decade. An indicator of the success of an account is the number of followers the account has. Therefore, various efforts have been made to increase the number of followers using fake accounts or fake users. Previous studies have proposed machine learning that can classify real and fake users. This study suggests identifying counterfeit users and using machine learning to accurately distinguish between actual and false users. To achieve higher accuracy outcomes in this study, the XGBoost method with hyperparameter tuning is applied. By applying hyperparameter tuning, the accuracy value of the XGBoost model can be raised once more, making it the model with the highest accuracy, with a value of 90.49%. This performance surpasses that of models developed for earlier research. Business owners have the choice of selecting the top influencers to hire for their services to advertise products based on the percentage of real followers from an account.