DIA-English-Arabic neural machine translation domain: sulfur industry
Creators
- 1. Department of Electrical Engineering, Shirqat Engineering College, Tikrit University, Tikrit, Iraq
- 2. Department of English Language, College of Basic Education, Tikrit University, Tikrit, Iraq
- 3. Department of Petroleum Control System, Oil Processing Engineering College, University of Tikrit, Tikrit, Iraq
- 4. Department of Civil Engineering, Engineering College, Tikrit University, Tikrit, Iraq
Description
The aim of this paper is the design and development a new English-Arabic neural machine translation (NMT) called DIA translation system. The main purpose of the designing system is to study translator limited sulfur industry domain as a stand-alone tool in order to improve the translation quality. Machine translation (MT) are very sensitive to the domains they were trained on and can be integrated with general (English-Arabic) MT systems. The proposed system has mainly four directions: supports chemical symbols, terms, phrase, and text and it is evaluated by using (1,200) various English declarative sentences which written by English language experts. The obtained results indicate that this system is high effective and has an accuracy of 79.33% in comparison with Google translator which has 38.67% for the same test samples.
Files
52 28049.pdf
Files
(336.5 kB)
Name | Size | Download all |
---|---|---|
md5:938987e5458c0f54679a7abe153ee64b
|
336.5 kB | Preview Download |