Multi-label Comment Classification Using GloVe-RNN Framework

Multi-label comment classification using GloVe embedding is addressed using deep neural networks in this paper. Online web media have opened a platform for users to express their thoughts on various contemporary issues and set no limit for their opinions. Unfortunately, the freedom of expression often leads to the usage of explicit language, which may hurt the readers. GloVe embedding vectors are computed in the front-end. Long short-term memory (LSTM), convolutional neural network (CNN), and gated recurrent neural network (GRU) models have been used for the classification phase. The performance is evaluated on the Kaggle comment classification dataset. The evaluation of results shows that recurrent neural network-based models outperform convolutional neural network-based models.