Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{178592,
        author = {BALIJA RAKESH and Akkamahadevi C and Allu Pravallika and Rahul Poldas and Ganesh Likith},
        title = {FAKE SOCIAL MEDIA PROFILE DETECTION AND REPORTING},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {12},
        pages = {5391-5397},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=178592},
        abstract = {This project presents a robust system for detecting fake social media profiles using a hybrid approach that leverages both classical machine learning and deep learning techniques. The system combines Logistic Regression for efficient baseline classification and Deep Neural Networks (DNNs) for capturing complex patterns in user behavior and content. It extracts a diverse range of features, including textual metadata, profile images, and user activity metrics, enabling comprehensive analysis. For image-based analysis, the system incorporates the Bag-of-Visual-Words (BoVW) model to transform profile pictures into visual histograms that feed into the classifier, enabling detection of reused or manipulated images often associated with fake accounts.
Additionally, text data from bios, usernames, and posts are processed using NLP techniques such as TF-IDF and sentiment analysis to identify linguistic patterns typical of bots or deceptive accounts. To manage scalability and maintain computational efficiency, the system uses gradient accumulation and batch processing during DNN training. The dataset is compiled from publicly available sources, consisting of real and fake profiles labeled manually or verified through third-party tools. Evaluation metrics such as accuracy, precision, recall, and F1-score are used to measure the system’s performance, with experimental results demonstrating high efficacy in fake profile identification across multiple platforms.
This project not only contributes to digital safety by flagging inauthentic users but also offers a scalable framework for platform-level integration to enhance trust and authenticity in online social interactions.},
        keywords = {Fake Profile Detection, Social Media Security, Logistic Regression, Deep Neural Networks (DNN), Bag-of-Visual-Words (BoVW), Natural Language Processing (NLP), Feature Engineering, Image Analysis, Sentiment Analysis, Gradient Accumulation, Batch Processing, Machine Learning Classification, Bot Detection, Online Identity Verification, Text Mining, Profile Image Classification, Multimodal Detection Systems, Trust and Safety in Social Media.},
        month = {May},
        }