Sentiment interaction and multi-graph perception with graph convolutional networks for aspect-based sentiment analysis

作者：

Highlights：

• A new solution is proposed to address the limitation of GCN models.

• A novel GCN with sentiment interaction and multi-graph perception is constructed.

• Four types of adjacency graph are generated to enrich aspect-centric dependencies.

• A multi-graph perception mechanism is constructed to capture specific dependencies.

摘要

•A new solution is proposed to address the limitation of GCN models.•A novel GCN with sentiment interaction and multi-graph perception is constructed.•Four types of adjacency graph are generated to enrich aspect-centric dependencies.•A multi-graph perception mechanism is constructed to capture specific dependencies.

论文关键词：Graph convolutional networks,Internal semantic correlations,Sentiment interaction relations,Multi-graph perception,Aspect-based sentiment analysis

论文评审过程：Received 10 February 2022, Revised 29 August 2022, Accepted 29 August 2022, Available online 5 September 2022, Version of Record 16 September 2022.

论文官网地址：https://doi.org/10.1016/j.knosys.2022.109840