Sentiment Analysis of Social Media Comments

With the development in technology and introduction of Web 2.0, Social Media Applications had a great role in connecting people from across the globe. The popular way of sharing opinions on the internet became the comment section. As every coin has two sides, misuse of the comments section was inevitable. Some comments contained hateful or extremist content against certain people or communities. These comments need to be deleted or prevented to maintain sanity on social forums. With the help of Sentiment Analysis technique, we can filter all the inappropriate comments, and flag the user. This solution is tested and running successfully on a Social Media Clone Application. The dataset used for training the model is an IMDb dataset. The data collected is then pre-processed, removing all the HTML tag elements. Then it is passed to the SVM based Machine Learning Model to train it. After the model has been trained, it is used to find whether the comment posted on the application is positive or negative. The comment, if passed all checks, is allowed to be posted. The negative comments are deleted and the user is flagged for being inappropriate. Thus this solution can be applied to multiple social forums and websites, contributing to a healthy social environment.