<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>Given the specialized nature of English text translation in academic settings and the frequent absence of reliable reference materials, translation processes often lack verifiable evidence, impacting both efficiency and quality. This paper addresses these challenges by first developing a basic syntactic error correction model that leverages the structural features of recurrent neural networks (RNNs) and gated recurrent unit (GRU) networks to establish a Seq2Seq syntactic error correction framework. To enhance this model, we incorporate an Attention mechanism into the Seq2Seq-based English grammar error correction model. This innovation enables the model to swiftly focus on segments most pertinent to the current context, thereby boosting operational efficiency. Subsequently, we create a college English text translation data corpus using Numerical Intelligence techniques to maintain grammatical accuracy within the corpus. Comparative analysis of the model training reveals that the Seq2Seq model with the Attention mechanism achieves an accuracy rate of 41.7%, which represents a 9.19% improvement over the basic model, underscoring its significant advantage. Furthermore, the average accuracy rate for grammatical error correction stands at 72.87%. A practical application analysis shows a minimal difference of only 0.05 points between the model’s grammar correction scores and those of human teachers. The corpus developed using this enhanced grammar error correction model scored 86 overall, outperforming other corpora. Therefore, the augmented Seq2Seq model with the Attention mechanism proves highly effective for developing English text translation corpora in collegiate environments.</p>
</abstract>

Given the specialized nature of English text translation in academic settings and the frequent absence of reliable reference materials, translation processes often lack verifiable evidence, impacting both efficiency and quality. This paper addresses these challenges by first developing a basic syntactic error correction model that leverages the structural features of recurrent neural networks (RNNs) and gated recurrent unit (GRU) networks to establish a Seq2Seq syntactic error correction framework. To enhance this model, we incorporate an Attention mechanism into the Seq2Seq-based English grammar error correction model. This innovation enables the model to swiftly focus on segments most pertinent to the current context, thereby boosting operational efficiency. Subsequently, we create a college English text translation data corpus using Numerical Intelligence techniques to maintain grammatical accuracy within the corpus. Comparative analysis of the model training reveals that the Seq2Seq model with the Attention mechanism achieves an accuracy rate of 41.7%, which represents a 9.19% improvement over the basic model, underscoring its significant advantage. Furthermore, the average accuracy rate for grammatical error correction stands at 72.87%. A practical application analysis shows a minimal difference of only 0.05 points between the model’s grammar correction scores and those of human teachers. The corpus developed using this enhanced grammar error correction model scored 86 overall, outperforming other corpora. Therefore, the augmented Seq2Seq model with the Attention mechanism proves highly effective for developing English text translation corpora in collegiate environments.

Construction of English Numerical Intelligence Text Translation Data Corpus in Colleges and Universities

Applied Mathematics and Nonlinear Sciences

This work is licensed under the Creative Commons Attribution 4.0 International License.

Given the specialized nature of English text translation in academic settings and the frequent absence of reliable reference materials, translation processes...

Construction of English Numerical Intelligence Text Translation Data Corpus in Colleges and Universities

Online veröffentlicht: 07. Juni 2024

Seitenbereich: -

Eingereicht: 27. Feb. 2024

Akzeptiert: 30. Apr. 2024

DOI: https://doi.org/10.2478/amns-2024-1442

Schlüsselwörter
Recurrent Neural Network, BiGRU Model, Attention Mechanism, Seq2Seq Model, Corpus Construction

© 2024 Xiang Zhai, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Construction of English Numerical Intelligence Text Translation Data Corpus in Colleges and Universities

Online veröffentlicht: 07. Juni 2024

Seitenbereich: -

Eingereicht: 27. Feb. 2024

Akzeptiert: 30. Apr. 2024

DOI: https://doi.org/10.2478/amns-2024-1442

SchlüsselwörterRecurrent Neural Network, BiGRU Model, Attention Mechanism, Seq2Seq Model, Corpus Construction

© 2024 Xiang Zhai, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Schlüsselwörter
Recurrent Neural Network, BiGRU Model, Attention Mechanism, Seq2Seq Model, Corpus Construction