Text sentiment analysis based on Glove model and United Network
Duojiao Li, Chenwan He, Ming Chen · Journal of Physics Conference Series · 2021
Abstract Traditional convolutional neural network has been proved to have good classification effect in text sentiment analysis experiments. However, CNN ignores the context dependent information of sentences, which affects the classification effect. Therefore, this paper proposes a text sentiment analysis model based on CapsuleNets and Attention mechanism. Firstly, the word vector is trained by using Glove model; secondly, the local features are extracted by using the features of double-layer capsule network; then, attention mechanism is added to assign the corresponding weight to the extracted text information; finally, we complete classification. The experimental results show that the proposed model can effectively improve the accuracy of emotion classification.