An Integrated CNN with LSTM for Sentiment Analysis to Detect User Emotion from Comments on Social Networks

Identifying users’ emotions through their short comments on social networks is an issue that has attracted a lot of attention. This problem belongs to the short text data classification problem direction based on many different models such as traditional, machine-learning, and deep-learning models. Convolutional neural networks (CNN) and recurrent neural networks (RNN) using LSTM techniques are recently commonly applied models for this problem with many outstanding advantages. Solving the emotion recognition problem with high performance requires a suitable model that can learn good data features. In this work, we propose a model called CNN_LSTM4SA that integrates CNN with LSTM technique, and word-embedding in text feature representation to identify user emotions based on a Vietnamese corpus. Experimental results are compared with many other related models to demonstrate the effectiveness of our proposed model.