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

<p>In this paper, we improve the performance of machine translation emotion recognition by fusing word features and linguistic features of environmentally friendly text with LSTM neural networks. Multi-channel LSTM’s unbalanced emotion classification method can address the problem of unbalanced numbers of samples in each emotion category in the emotion classification task. The bilingual semantic similarity degree feature is added to the translation process of English environmental protection data so as to make the translation of English environmental protection data more accurate. The proposed method is validated through experiments, and the results show that the translation quality of the LSTM model is generally higher than that of other models, and the mean value of the translation quality of the LSTM on the test17 dataset is 50.4. In terms of environmental protection terminology, mistranslations, common word mistranslations, etc. There is a significant reduction of 2–3 times in the absolute number of the LSTM model compared to the traditional method. A total of 94,788 corpora were obtained from 10 environmental translations, such as Silent Spring and Solace of the Heart, and 2,311 favorite emotion words were identified. Comparative analysis of the samples reveals that most of the translation results of some environmental terms are in the range of 0.7–0.9, and the range of semantic overlay between the machine translation results of the LSTM model and the standard translations is large. Most users believe the model improves the translator’s translation level and diction accuracy.</p>
</abstract>

In this paper, we improve the performance of machine translation emotion recognition by fusing word features and linguistic features of environmentally friendly text with LSTM neural networks. Multi-channel LSTM’s unbalanced emotion classification method can address the problem of unbalanced numbers of samples in each emotion category in the emotion classification task. The bilingual semantic similarity degree feature is added to the translation process of English environmental protection data so as to make the translation of English environmental protection data more accurate. The proposed method is validated through experiments, and the results show that the translation quality of the LSTM model is generally higher than that of other models, and the mean value of the translation quality of the LSTM on the test17 dataset is 50.4. In terms of environmental protection terminology, mistranslations, common word mistranslations, etc. There is a significant reduction of 2–3 times in the absolute number of the LSTM model compared to the traditional method. A total of 94,788 corpora were obtained from 10 environmental translations, such as Silent Spring and Solace of the Heart, and 2,311 favorite emotion words were identified. Comparative analysis of the samples reveals that most of the translation results of some environmental terms are in the range of 0.7–0.9, and the range of semantic overlay between the machine translation results of the LSTM model and the standard translations is large. Most users believe the model improves the translator’s translation level and diction accuracy.

Exploring the Direction of English Translation of Environmental Articles Based on Emotional Artificial Intelligence Learning Models

Applied Mathematics and Nonlinear Sciences

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

In this paper, we improve the performance of machine translation emotion recognition by fusing word features and linguistic features of environmentally...

Exploring the Direction of English Translation of Environmental Articles Based on Emotional Artificial Intelligence Learning Models

Publicado en línea: 02 jul 2024

Páginas: -

Recibido: 21 feb 2024

Aceptado: 20 may 2024

DOI: https://doi.org/10.2478/amns-2024-1523

Palabras clave
LSTM model, Environmental data, Weight matrix, Emotion recognition, Emotion classification

© 2024 Wensen Xian, published by Sciendo

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

Exploring the Direction of English Translation of Environmental Articles Based on Emotional Artificial Intelligence Learning Models

Publicado en línea: 02 jul 2024

Páginas: -

Recibido: 21 feb 2024

Aceptado: 20 may 2024

DOI: https://doi.org/10.2478/amns-2024-1523

Palabras claveLSTM model, Environmental data, Weight matrix, Emotion recognition, Emotion classification

© 2024 Wensen Xian, published by Sciendo

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

Palabras clave
LSTM model, Environmental data, Weight matrix, Emotion recognition, Emotion classification