Block-Level Dependency Syntax Based Model for End-To-End Aspect-Based Sentiment Analysis

End-to-End aspect-based sentiment analysis (E2E-ABSA) aims to jointly extract aspect terms and identify their sentiment polarities. It would be helpful to utilize block-level inner-relation and inter-relation of aspect terms and opinion terms, which is often ignored in most previous syntax-based ABSA models. This paper proposes a block-level dependency syntax parsing (BDEP) based model to enhance the performance of E2E-ABSA. BDEP is first built to model the aspect-oriented block relations based on routine dependency syntax and part-of-speech. Then the BDEP-guided interactive attention (BDEP-IAM) module and the adaptive semantic-syntactic fusion module are leveraged to jointly extract aspect terms and identify their sentiment polarities. Experiments on four benchmark datasets are conducted and the results show that our proposed model outperforms the other compared state-of-the-art (SOTA) methods in all four datasets.