Automated Tool for Toxic Comments Identification on Live Streaming YouTube

The necessity for content moderation on social media websites is increasing every day. The reason behind this is the anonymity of an individual which is provided by the internet and that they can exercise on any streaming website like YouTube. This project aids the creators/moderators in maintaining and stabilizing the toxicity of comments on their channel/page. A moderator has the authority to delete/hide any inappropriate comment posted by any user. During the pandemic of Covid-19, a significant, large user base immigrated to social media websites particularly streaming websites like YouTube. This surge in users resulted in an increased need for moderators. Any user can post any hateful/toxic or obscene comment that moderators can miss due to the huge volume of comments. Using NLP (Natural Language Processing), a model can be implemented directly onto any live-streaming chat session which can identify any toxic/obscene or threat comment and can flag them under the respective category. This model is real time and formulated using NLP techniques that are used in order to achieve the task.