Ahierarchical neural model for target-based sentiment analysis

摘要

A convolutional neural network-regional long Short-Term memory (CNN-RLSTM) is proposed, which is a convolutional neural network-regional long short-term memory (CNN-RLSTM) that combines CNN and regional LSTM. The model can effectively distinguish the affective polarity of different targets through a regional LSTM while reducing the training time of the model. In addition, the model can retain the sentiment information of the whole sentence through a CNN network at the sentence level. Experimental results on different data sets show that the CNN-RLSTM model is better than the traditional model and the deep network model.