FAKE PROFILES DETECTION FROM FACEBOOK INSTAGRAM AND TWITTER USING ARTIFICIAL INTELLIGENCE

Abstract

The proliferation of social media platforms since the early 2000s has transformed communication, commerce, and information dissemination globally. However, this growth has been accompanied by the emergence of fake profiles, which undermine user trust and facilitate malicious activities such as misinformation, fraud, and cyberbullying. According to a 2023 report by Statista, over 5% of social media accounts across major platforms like Facebook, Instagram, and Twitter are estimated to be fake. Addressing this issue necessitates advanced detection mechanisms leveraging artificial intelligence. This project employs machine learning and deep learning models, trained on datasets sourced from Kaggle, to identify fraudulent profiles with high accuracy. Models such as Artificial Neural Networks (ANN)-98.34%, K-Nearest Neighbors (KNN)-98.24%, advaned Light Gradient Boosting Machine (LGBM)-99.44%, Logistic Regression (LR)-97.24%, Random Forest-96.88%, and Support Vector Machines (SVM)-96.66%. The system allows users to detect fake profiles using profile and location details, enhancing the robustness of identification. Additionally, the web application facilitates administrative functionalities including dataset management, algorithm execution, and user feedback analysis. By integrating historical data trends and current statistical insights, the developed solution offers a comprehensive approach to mitigating the risks posed by fake social media profiles, thereby fostering a safer online environment.