Fake profiles are often created to spread misinformation, disinformation, and propaganda. Detecting and removing these profiles is crucial to curb the dissemination of false information and maintain the integrity of information shared on social media. In this study, we present an enhanced algorithm for the detection of fake social media profiles, utilizing machine learning techniques such as Gradient Boosting, Random Forest, and Support Vector Machine. The algorithm incorporates a range of profile features, including the presence of a profile picture, characteristics of the username and full name (length, numbers, equality), description length, external URL presence, account privacy, and key metrics like the number of posts, followers, and follows. The primary objective is to address the escalating issue of fraudulent activities and misinformation on social media platforms. The proposed algorithm leverages ensemble learning to improve the accuracy and reliability of identifying deceptive profiles. Additionally, we introduce a Flask-based web application to deploy the Random Forest algorithm, enabling real-time detection and providing a user-friendly interface. To evaluate the algorithm\\\'s performance, precision, recall, and F1 score are employed as key metrics. Precision measures the accuracy of positive predictions, recall gauges the algorithm\\\'s ability to capture all positive instances, and the F1 score balances precision and recall. Through comprehensive testing and validation, our algorithm aims to contribute to the advancement of online security, fostering user trust and mitigating the impact of fake profiles in the dynamic landscape of social media.
Pooja Dharshini K (Thu,) studied this question.