.png){kind=icon}
.png){kind=icon}
Return
HSAN-capsule: A novel text classification model
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2021.12.064.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Most text classification methods fail to fully consider the important role of the hierarchical structure of the text in determining the text category, and cannot well extract enough text sequence information. To address this problem, a novel model based on hierarchical self-attention mechanism capsule network was proposed for text classification, which was composed of two components, i.e., the hierarchical self-attention network and capsule network. First input the text data processed by word embedding into the hierarchical self-attention network for feature extraction, model the text from the two levels of words and sentences to extract the hierarchical features of the text; then the results are passed into the capsule network in order to refine the relationship between the part of the text and the whole, and further extract richer text semantic information. The experimental results on 5 text classification datasets show that compared with other baseline models, the model in this paper has achieved better classification results.(c) 2022 Elsevier B.V. All rights reserved.
Keywords:
Text classification
Hierarchical network
Recurrent neural network
Capsule network
Self-attention mechanism
Journal
IF:
6.5
Papers:
2.5W
Citations:
6.5W