Semantic Relatedness Enhanced Graph Network for aspect category sentiment analysis

作者：

Highlights：

• A hierarchical approach is adopted to deal with aspect category sentiment analysis.

• A semantic graph is built on semantic similarity and relatedness measures.

• A novel Edge-gated GCN is proposed to explore multi-channel edge features.

• A new sub-objective loss function of semantic information is considered.

摘要

•A hierarchical approach is adopted to deal with aspect category sentiment analysis.•A semantic graph is built on semantic similarity and relatedness measures.•A novel Edge-gated GCN is proposed to explore multi-channel edge features.•A new sub-objective loss function of semantic information is considered.

论文关键词：Aspect category sentiment analysis,Semantic relatedness,Edge-Gated Graph Convolutional Network,Aspect–context attention

论文评审过程：Received 17 May 2021, Revised 26 August 2021, Accepted 16 January 2022, Available online 5 February 2022, Version of Record 9 February 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.116560