Hiding Expletive Comments in Mobile Applications using CNN and LSTM based NLP Classification Models

Often, e-commerce mobile applications show the public comments of consumers about the products they sell. Sometimes, these comments contain foul language, which is inappropriate to be shown on the public platform. App developers would want to hide them and show them only after the consent from the app user. The goal of this study is to find the optimal way to classify comments as expletive or not using the NLP classification model. This study utilized CNN and LSTM algorithms to train the expletive language classification model. These models are used by the mobile application to find whether comments from users are expletive in nature or not. If a comment is found to be expletive, it will be hidden. The mobile app will also provide an option to unhide the expletive comments if the user wants to see them. LSTM models are found to be more accurate than CNN models with large datasets. Hiding expletive comments is very important for organizations to meet the guidelines of various countries. Deep learning provides an accurate and novel approach to achieve this feature.