Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2025
|
| Online Access: | https://doi.org/10.5281/zenodo.17092909 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Table of Contents:
- <p><span lang="EN-US">Accurate and context-sensitive language translation remains a persistent challenge in natural language processing, particularly for morphologically and semantically rich languages such as French. This study introduces a hybrid Bidirectional LSTM–GRU model for English-to-French translation, designed to integrate the representational advantages of both architectures to improve performance. The methodology encompassed preprocessing a bilingual corpus using tokenization and padding, constructing a stacked hybrid architecture with bidirectional layers, and employing optimization strategies including EarlyStopping and ModelCheckpoint. The proposed model attained an accuracy of 96.51% with a training loss of 0.1035, demonstrating its ability to handle sentences of up to 55 tokens without performance degradation. The results highlight the robustness of the hybrid approach, while its deployment as a SaaS application underscores its scalability and applicability to real-time machine translation tasks.</span></p>