Answers are grounded in this paper and cite their sources.
This research demonstrates improved classification accuracy using a Transformer-LSTM model, indicating enhanced semantic understanding and discrimination capabilities.
With the rapid development of natural language processing technology, the importance of English word sense comparison classification has become increasingly prominent in areas such as semantic understanding and machine translation. The research results in this field play a crucial role in enhancing the performance of language processing systems. However, most existing English word sense comparison classification methods currently have significant shortcomings when dealing with complex semantic relationships. They struggle to accurately capture long-distance dependencies between words and exhibit weak discrimination capabilities when handling polysemy and synonyms, posing challenges in building high-performance comparison classification systems. The limitations of existing methods are mainly reflected in two aspects: on one hand, traditional models find it difficult to effectively extract semantic features when processing long sentences and complex semantic structures; on the other hand, a single model cannot simultaneously account for sequence information retention and long-distance dependency capture, leading to limited classification accuracy. This research focuses on these issues by proposing the construction of an English word sense comparison classification system using Transformer and LSTM. The self-attention mechanism of the Transformer can effectively capture long-distance dependencies between words, enhancing the ability to extract semantic features; the memory functionality of LSTM preserves sequence information, and the combination of the two optimizes classification performance. In our experiments, we constructed an English word sense comparison dataset containing 10,000-word pairs and used cross-validation for evaluation. The results show that, compared to the classification accuracy of the single Transformer model at 80.2% and the LSTM model at 78.9%, the Transformer LSTM model improved the classification accuracy to 85.6%, achieving an average percentage improvement of 5.4% and 6.7%. In the comparison tasks dealing with polysemy and synonyms, the model reached accuracies of 88.3% and 86.1%, respectively, demonstrating excellent discrimination capabilities and stable performance in handling complex semantic relationships and long sentences, significantly outperforming traditional models.
No takes yet. Share an insight, caveat, or question.
Tan et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: