Error Analysis Of Machine Translation In Management Text

With the development of computer and internet technology, machine translation research has entered a new stage, and the research results play an important role in information communication among different languages all over the world. The quality of current machine translation is not desirable, but it has provided convenient services in restricted language translation. Using two management papers as analytic texts, the thesis compares the machine translation texts generated by Transn and Systran system with manual translation texts, and describes the typical errors of machine translation in terms of terminology, syntax and symbols in management texts. Based on the error analysis, the paper aims to make supplements for the rules of machine translation systems with regard to term bases, part-of-speech tagging and deep syntactic analysis. Therefore, the quality of machine translation in restricted languages, management text in particular, could be improved.