Aspect sentiment analysis with heterogeneous graph neural networks

作者:

Highlights:

• The interactive aspect words and contexts are employed for the sentence encoder and conduct parameters sharing.

• The syntactic sentence relations, prior sentiment dictionary and some part-of-speech tagging information, are fused in the heterogeneous graph.

• Experimental results show that our approach can outperform most state-of-the-art aspect sentiment classification methods.

摘要

•The interactive aspect words and contexts are employed for the sentence encoder and conduct parameters sharing.•The syntactic sentence relations, prior sentiment dictionary and some part-of-speech tagging information, are fused in the heterogeneous graph.•Experimental results show that our approach can outperform most state-of-the-art aspect sentiment classification methods.

论文关键词:00-01,99-00,Graph attention networks,Aspect sentiment analysis,Opinion mining,Heterogeneous graph neural network,Multi-head attention mechanism,Graph convolution neural networks

论文评审过程:Received 10 December 2021, Revised 24 March 2022, Accepted 21 April 2022, Available online 13 May 2022, Version of Record 13 May 2022.

论文官网地址:https://doi.org/10.1016/j.ipm.2022.102953