Aspect-based Sentiment Analysis of Reviews for Requirements Elicitation (Defence)

Fine-grained sentiment analysis faces ongoing challenges in Aspect Sentiment Triple Extraction (ASTE), particularly in accurately capturing the relationships between aspects, opinions, and sentiment polarities. While researchers have made progress using BERT and Graph Neural Networks, the full potential of advanced language models in understanding complex language patterns remains unexplored. We introduce DESS (DeBERTa Enhanced Syntactic Semantic), a new approach that builds upon previous work by integrating DeBERTa’s enhanced attention mechanism to understand context and relationships in text better. Our framework maintains a dual-channel structure, where DeBERTa works alongside an LSTM channel to process both meaning and grammatical patterns in text. We’ve carefully refined how these components work together, paying special attention to how different types of language information interact. When we tested DESS on standard datasets, it showed meaningful improvements over current methods, with F1-score increases of 4.85, 8.36, and 2.42 in identifying aspect opinion pairs and determining sentiment accurately.Building on DESS, we further introduce SGESS (Span Level Graph Enhanced Syntactic Semantic), which adds a novel Span Level Graph Reasoning module that enables candidate entity spans to communicate with each other before sentiment classification. The Span Level Graph constructs dynamic graphs over candidate spans, applies GNN message passing with gated fusion, and achieves state of the art results on all four SemEval benchmarks: 80.75% F1 on 14RES, 80.01% on 16RES, 77.19% on 15RES, and 72.42% on 14LAP. Looking deeper into the results, we found that DeBERTa’s sophisticated attention system helps DESS handle complicated sentence structures better, especially when important words are far apart., and the Span Level Graph provides consistent improvements of +2.8 to +3.6 F1 points by letting aspect and opinion spans share information before the final sentiment decision. Our findings suggest that upgrading to more advanced language models, combined with inter span communication mechanisms, when thoughtfully integrated, can lead to real improvements in how well we can analyze sentiments in text.