DeTox: A WebApp for Toxic Comment Detection and Moderation

• Published in: 2024 International Conference on Trends in Quantum Computing and Emerging Business Technologies pp. 1 - 5
• Main Authors: N, Prudhvish, G, Nagarajan, U, Bharath Kumar, Vardhan B, Harsha, L, Tharun Kumar
• Format: Conference Proceeding
• Language: English
• Published: IEEE 22.03.2024
• Subjects: API; BERT; Machine learning; Real-time systems; Solids; Toxic comments; User interfaces; Video on demand; Web page design; Web pages; YOUTUBE
• DOI: 10.1109/TQCEBT59414.2024.10545229

The extensive adoption of internet platforms such as YouTube has transformed communication and information exchange, compelling people to share their opinions and participate in global conversations. Open communication can, however, also encourage the spread of offensive material, such as remarks that are derogatory or involve threats or hate speech. Such offensive remarks have the potential to damage users' mental health by fostering a hostile and unsafe environment that discourages meaningful relationships. We introduce DeTox, a web application that uses machine learning techniques to detect and eliminate harmful comments from YouTube videos in order to address this problem. For the purpose of classifying comments, DeTox uses ML and DL models, which guarantees precise identification of harmful content. The YouTube Data API is integrated by the system to retrieve comments from specific videos and eliminate any harmful remarks found. A detailed examination of the Toxic Comment Classification Challenge dataset, made available by Kaggle, was necessary for the development of DeTox. To find common patterns in hazardous language, preprocessing and examination of the data were done in order to examine the distribution of toxic and non-toxic remarks. FastAPI is a high-level Python web framework that makes web application development easier and is used by the DeTox online application. The application includes a user friendly interface for creating accounts, logging in, and selecting videos to moderate. The application also includes a user interface for reviewing toxic comments and choosing to remove or keep them. DeTox is a valuable tool for moderating comments on YouTube videos. The application utilizes a machine learning model to accurately identify toxic comments, and it provides a vi user-friendly interface for removing or keeping the comments. DeTox has the potential to make YouTube a more friendly and safe environment for users.