Dual-Enhanced Dependency Tree-Based Sentiment Analysis Model

Traditional aspect-level sentiment analysis models typically employ unidirectional contextual interaction mechanisms, whereas actual language comprehension requires bidirectional semantic interactions. To address this, this study proposes a Syntax-based Interactive Graph Attention model (An Dual-Enhanced Dependency Tree-Based Sentiment Analysis Model, hereafter referred to as the DTSA model) for aspect-level sentiment analysis, which constructs a bidirectional dynamic interaction mechanism to effectively integrate syntactic structural features with contextual semantic information. Experimental results on the standard Restaurant and Laptop datasets demonstrate that the DTSA model significantly outperforms baseline models in both classification accuracy(Acc) and F1-score(F1) metrics, validating the rationality of its design and its performance advantages.