Research on Stylistic Features in Translation Based on Supervised Learning Algorithms
- Resource Type
- Conference
- Authors
- Zheng, Jianyu; Sun, Jin; Jiang, Yanting; Zhu, Yun
- Source
- 2018 IEEE 4th International Conference on Computer and Communications (ICCC) Computer and Communications (ICCC), 2018 IEEE 4th International Conference on. :2332-2337 Dec, 2018
- Subject
- Communication, Networking and Broadcast Technologies
Computing and Processing
Classification algorithms
Support vector machines
Linguistics
Training
Supervised learning
Testing
Machine learning algorithms
supervised learning
text style
classification algorithm
stylistic features
information gain
comparative linguistics
- Language
This paper uses supervised classification algorithms to study the changes of stylistic features caused by the translators and the differences between Chinese and English languages in the process of translation. We collect the original and translation works of A Tale of Two Cities and Jane Eyre in English and Chinese. Firstly, we make a list of features and calculate these values of texts for each stylistic feature. Then the Information Gain (IG) values of each stylistic feature in both Chinese and English texts are calculated respectively. Finally, the values of each feature are added to each classifier in order according to their IG values to observe the performance of each classifier. This paper analyzes the change about stylistic features' competence in distinguishing texts in the process of translation, from the perspective of contrastive linguistics and the translator roles and so on. The experimental results show that the ability of each classifier to distinguish the style differences in the translation version is mostly decreased compared with the original one. After translation, the IG values and orders of each stylistic feature vary greatly.