Text Sentiment Classification Based on Attention Mechanism and Decomposition Convolutional Neural Network model
| Autor: | Xu Zhao, Yujie Huang, Haibo Chen, Dong Cao, Hui Li, Yunbin Fu |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Structure (mathematical logic)
0209 industrial biotechnology Artificial neural network business.industry Mechanism (biology) Computer science Computer Science::Computation and Language (Computational Linguistics and Natural Language and Speech Processing) Pattern recognition 02 engineering and technology Convolutional neural network 020901 industrial engineering & automation Simple (abstract algebra) 0202 electrical engineering electronic engineering information engineering Decomposition (computer science) Feature (machine learning) 020201 artificial intelligence & image processing Artificial intelligence Layer (object-oriented design) business |
| Zdroj: | 2020 IEEE International Conference on Advances in Electrical Engineering and Computer Applications( AEECA). |
| Popis: | In view of the simple structure of a single neural network model, the traditional convolutional neural network cannot fully extract deep text features. This paper proposes a text sentiment classification based on Attention Mechanism and Decomposition Convolutional Neural Network model. First, comprehensive text features are obtained with the help of parallel Decomposition Convolutional Neural Network (DCNN). Secondly, it combines the attention mechanism to extract important feature information and improve the optimized text sentiment classification effect. Finally, the text is sentimentally classified at the classification layer. After conducting multiple sets of comparative experiments on three sets of Chinese data sets, the accuracy of the model in this paper reached 92.06%, 91.08% and 92.71% respectively. It is proved that the proposed model performs better than the single-channel model, and the use of Decomposed Convolutional Neural Network is better than traditional Convolutional Neural Network. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|