A Novel Aspect-based Sentiment Analysis Network Model Based on Multilingual Hierarchy in Online Social Network

摘要

In recent years, sentiment analysis based on aspects has become one of the research hotspots in thefield of natural language processing. Aiming at the fact that the existing network model cannotfully obtain the interrelationship between sentences in the same comment and the long-distancedependence of specific aspects in the whole comment, a multilingual deep hierarchical model combiningregional convolutional neural network and bidirectional LSTM network is proposed. Themodel obtains the time series relationship of different sentences in the comments through theregional CNN, and obtains the local features of the specific aspects in the sentence and the longdistancedependence in the whole comment through the hierarchical attention network. In addition,the model improves the word vector representation based on the gate mechanism to makethe model completely independent of the language. Experimental results for different domain datasetsin multi-language show that the proposed model achieves better classification results than thetraditional deep network model, the network model combining with the attention mechanism andconsidering the relationship between sentences.