Author(s)

Zheng Jia

Corresponding Author

Zheng Jia

ABSTRACT

Emotional recognition has a great influence on people's daily behavioral interactions. With the emergence of massive data, especially various digital social media, newspapers, magazines, books, daily conversations and so on. Identifying emotion of text automatically by computer, which is of great significance to the user's sentiment analysis, network public opinion monitoring and other fields. In recent years, with the rise of deep learning, great success has been achieved in many fields, especially the emergence of gated recurrent neural networks, which are good at processing sequence data such as text and solving the problem of Long-Term dependencies. Based on this, this study proposes a multi-label emotion recognition of text based on Bi-directional Long Short-Term Memory and Word embedding technology. The experimental results show that the performance of our neural network model outperformed other neural network models. The training set, the validation set and the test set have an accuracy of 0.753, 0.6228, and 0.6136, respectively, which means that the state-of-the-art 7 kinds of emotion recognition for the text are realized, namely, happiness, anger, sadness, disgust, surprise, fear, normal

KEYWORDS

Emotion recognition of text, Deep Learning (DL), Bi-directional Long Short-Term Memory (BiLSTM), Word Embedding, 7 kinds of emotions