Abstract:
The wordlevel shallow convolutional neural network (CNN) model has achieved good performance in text classification tasks. However, shallow CNN models cant capture longrange dependencies, which affects the model's performance in text classification tasks, but simply deepening the number of layers of the model does not improve the models performance. This paper proposes a new wordlevel text classification model WordCNNAtt, which uses CNN to capture local features and position information, and captures longrange dependencies with selfattention mechanism. The accuracy of WordCNNAtt, on the five public datasets of AGNews, DBPedia, Yelp Review Polarity, Yelp Review Full, Yahoo! Answers, is 0.9%, 0.2%, 0.5%, 2.1%, and 2.0% higher than the wordlevel shallow CNN model respectively.