Silencing Hate: Developing a Malignant Comment Filter for Safer Social Media Spaces

The proposed application aims to provide social media platforms with an efficient tool to filter out malignant comments and hate speech that is abusive in nature, allowing users to either censor or hide these comments. The dataset used for this project is obtained from a Kaggle competition, containing approximately 220K samples of text/comments posted by Wikipedia commenters labeled based on their type of toxicity. The project intends to use state-of-the-art models such as LSTM and Transformers to improve the performance of the application. The evaluation of the models will be based on precision and recall, with the goal of minimizing false negatives while minimizing false positives to maintain user experience.