A team of researchers from the International Institute of Information Technology (IIIT), Naya Raipur, is working on a unique research project that would work on blocking negative comments from social media. The project, called ‘Comment Toxicity Prediction using Deep Learning’ is expected to play an important role to keep the users safe and to maintain positivity in social media.
The research team comprises three students, namely Sayantan Roy, Rishi Shounak, and Vivek Kumar, who worked under the guidance of Vivek Tiwari and made it possible.
Through this project, the students have worked on software that will work as a filter for social media. The software has its own data based on which it will decide whether the comment is negative or positive and if anyone posts a negative comment, the filter will automatically delete it. This software will work on all social media platforms including Facebook, Instagram, Twitter, and YouTube. This software will also help to improve the productivity of social media.
“In today’s time, such software is the need of the hour because people have turned insensitive and are using social media for hurting other people,” said Tiwari. People also use social media for breaching government rules so the software will help maintain balance on social media. He also added that they will also try for an international patent.
The software works on a machine learning model. In this, the software checks the comments through the machine learning process. The machine after going through the comment will score the comment, if the score is negative, the comment will be deleted automatically and if the comment is positive, the comment will remain posted on the wall. The software had its data based on which the software guessed the toxicity of the comment. The software error is 2.0 mean error. At present, the students for the testing purpose, have prepared a web app for this software and anyone can easily download the link and can scrape all the negative comments from their social media. The software allows the user to decide first whether he wants to see the comment or wants to delete it immediately.
One of the students, Rishi Shounak, informed that there are still some issues in the software as it fails to recognise sarcastic comments and he added that their team was working hard on this and will soon overcome it. He said that they will add more data sets to improve the problem.