摘要
针对传统翻译算法对复杂长句切分翻译时翻译结果不够精准、回收率较高的问题,研究一种基于语义特征的复杂长句切分式翻译算法。该算法依照汉语短语类别进行语句关系判定;提取嵌套式以及非嵌套式句子中的特定词语明确复杂长句的语义特征;根据整句或分句中的标点符号和特定连词进行句子切分;建立语义感知单元和语义特征单元组成的预处理模块对复杂长句逐句翻译。实验结果表明,该翻译算法与传统翻译算法相比,结果准确率高出6.65%,平均回收率低了6.29%。该算法更能满足实际工作需要。
Aiming at the problem that the traditional translation algorithm is not accurate enough and has a high recovery rate when translating complex long sentences,a novel segmentation translation algorithm based on semantic features is studied.The algorithm determines sentence relations according to Chinese phrase categories.Extracting specific words in nested and non-nested sentences to identify the semantic features of complex long sentences;According to the punctuation marks and specific conjunctions in the whole sentence or clause,sentence segmentation is carried out.A preprocessing module composed of semantic perception unit and semantic feature unit is established to translate complex long sentences sentence by sentence.Experimental results show that compared with the traditional translation algorithm,the accuracy of the results is 6.65%higher than the traditional algorithm,and the average recovery rate is 6.29%lower than the traditional algorithm.It can be seen that the designed algorithm can better meet the actual work needs.
作者
饶岩岩
RAO Yanyan(College of Foreign Language of Longyan University,Longyan 364012,China)
出处
《周口师范学院学报》
CAS
2020年第1期95-99,共5页
Journal of Zhoukou Normal University
基金
福建省教育厅2018年中青年教师教育科研项目“客家文化外宣翻译中的跨文化意识研究”(JAS180501)。
关键词
复杂长句
词语关系
语义特征
长句切分
预处理模块翻译
切分式翻译算法
complex long sentences
word relations
semantic features
long sentence syncopation
preprocessing module translation
tangent translation algorithm
