关键词: 方面情感三元组抽取, 句法依赖关系, 词性信息, 句法依赖距离, 图卷积网络

Abstract: In Natural Language Processing (NLP) tasks, Aspect Sentiment Triplet Extraction (ASTE) aims to identify therelationships among aspect terms, opinion terms, and sentiment polarity in text, serving as a crucial step in achieving fine-grained sentiment analysis. In current mainstream approaches, end-to-end models generally suffer from insufficient understanding of linguistic features and poor handling of the sparsity in sentiment expressions, which limits their accuracy and robustness. Pipeline models are prone to error propagation. To address these issues, Multi-View Language Feature-based Sentiment Lexicon (MVLF-SL) model was proposed. Multi-view linguistic features were utilized to enhance the model’s ability to understand context and implicit semantics, while a sentiment lexicon provided additional prior knowledge of sentiment. First, GraphConvolutional Networks (GCN) were used to represent multi-view linguistic features and obtain enhanced representations. Then, a dynamic fusion strategy was adopted to integrate the enhanced language features with sentiment lexicon information. After that, multi-layer GCNs were employed to further enhance the feature representations of aspect and opinion terms by incorporating adjacency relations and node features. Finally, a Boundary-DrivenTable-Filling (BDTF) method, improved with a biaffine attention mechanism, was used for decoding and extracting sentimenttriplets. Experimental results on four subsets (14res,14lap, 15res, and 16res) of the ASTE-DATA-V2 dataset show that, compared with the BDTF model, the F1 scores were improved by 0.57, 2.08, 2.20, and 1.93 percentage points, respectively. Theseresults indicate that the proposed model achieves better performance in aspect sentiment triplet extraction byeffectively leveraging linguistic features and external sentiment knowledge.

Key words: Aspect Sentiment Triplet Extraction(ASTE), syntactic dependency relations, Part-of-Speech Information, syntactic dependency distance, Graph Convolutional Network(GCN)