Error Types Of Machine Translated Patent Texts And Their Respective Post-Editing Strategies

This report is based on the translation project of Translation of Patent Literatures in Key Technology Fields.The project was initiated and commissioned by China Association for Science and Technology(CAST),and undertaken by Jiangsu Translators Association of Science and Technology(JTAST).For higher efficiency,the JTAST used the machine translation(MT)system of Jiangsu Sunyu Information Co.,Ltd.to convert the format of patents from PDF to Word and to translate them into Chinese automatically.Then,JTAST assigned the processed texts to translators for post-editing.In this project,the author was responsible for the work of post-editing.And in this thesis,the author analyzed the machine translated texts in translation practice.And the error types were generally divided into three categories of lexical errors,syntactic errors,and symbol errors.Furthermore,lexical errors includes term errors part of speech errors and missing words,syntactic errors contains syntactic element ordering errors and complex sentence structure errors,and symbol errors cover conversion errors and translation errors.For these error types,the author proposed respective post-editing strategies.This thesis proposed effective post-editing strategies according to the analysis of the error types of machine translated patents,and applied each strategy to correct errors in translation practice,hoping to shed light on how to improve the post-editing in machine translation with special reference to patent translation.