Fake Profile Detection in Instagram

Abstract

Abstract: In today's society, online communities have become an integral part of daily life, with individuals actively engaging in social media interactions regardless of their location or schedule. However, the prevalence of pseudo personas has contributed to sophisticated persistent attacks and other malicious activities, raising concerns about the privacy of personal data among online community members. This research seeks to identify and address the pressing issue of fraudulent identity projection on social media platforms, particularly focusing on Instagram. By employing automated techniques such as predictive modeling and image identification combined with text analysis, the study aims to enhance the detection of fake profiles. The research utilizes the chi-square method for feature selection and applies learner-centered algorithms like Random Forest (RF) and logistic regression for categorization. The evaluation of outcomes will be based on metrics including relevance, specificity, recall, f1- value efficiency, and accuracy. The exponential growth and influence of social media globally have underscored the urgency of mitigating the proliferation of fraudulent identities, which pose challenges ranging from propaganda to racial profiling. The findings of this research are expected to contribute to the development of effective strategies for identifying and addressing fake profiles, thus safeguarding the integrity and security of online communities