Combined machine translation method by utilizing segmentation technique

Abstract

The invention provides a combined machine translation method by utilizing a segmentation technique, in particular to a combined machine translation method for translating Chinese into English by utilizing a segmentation technique. Through reasonable segmentation of Chinese sentences, similarity calculation and an English generation rule, English text meeting requirements is obtained without dependence on a network database at a certain degree, and by only processing the Chinese sentences to be translated and combining with a set English translation rule at the same time, an accurate translation result can be obtained.

Description

technical field [0001] The invention belongs to the field of automatic translation, and in particular relates to a compound machine translation method using segmentation technology. Background technique [0002] With the development of smart devices, smart operating systems are becoming more and more diverse, such as Apple's IOS, Google's Android, Firefox's Firefox OS, etc., and smart devices integrated with these systems have also begun to be used more and more. Many users use these devices for daily activities such as gaming, socializing, reading and so on. [0003] As the society becomes more and more open, people now have more opportunities to read content that does not belong to their native language. Whether it is a hobby of reading, or because of professional study, work needs, etc., they often come across a lot of foreign language materials. At present, the most common method of querying foreign language vocabulary on smart devices is for the user to manually open t...

AI Technical Summary

Problems solved by technology

[0003] As the society becomes more and more open, people now have more opportunities to read content that does not belong to their native language. Whether it is a hobby of reading, or because of professional study, work needs, etc., they often come across a lot of foreign language materials. At present, the most common method of querying foreign language vocabulary on smart devices is that the user manually opens the foreign language query application and manually enters the word query. Some applications that do a little better, such as Youdao Dictionary

There are three main types of popular automatic translation methods at present. The first type is based on words, which use words as the basic unit of translation, regardless of context information and human linguistic knowledge. When translating, first find the target corresponding to each source language word language words, then insert and delete target language words, adjust their order, and combine them into target language sentences. The feature is that the translation is fast, but the accuracy is poor. The second type is phrase-based translation, and the translation granularity extends from words to Phrases, which better solve the problem of local context dependence, greatly improve the fluency and accuracy of translation. The third type is syntax-based translation, which introduces syntactic structure information into the translation process, but needs to introduce grammatical structure knowledge, and needs to be translated before translation. Use syntactic knowledge to adjust the word order of the source language, and use syntactic knowledge to reorder words after translation

[0004] At present, in the existing automatic machine translation, the third type of translation is the trend. However, in order to obtain a better translation effect, it is best to obtain the grammatical structure through online networking. In addition, the translation speed is relatively slow

Although the Internet has been widely used, however, with the change of the environment and the emergence of various temporary conditions, our smart devices cannot keep online all the time. Therefore, there is an urgent need for a composite The traditional machine translation method can get more accurate translation results when it is separated from the network as much as possible

Images