Intermediate Language

The task of a translator requires a great talent. The language consists of morphology, syntax and semantics, and must also consider the style and pragmatics. The wide range that can be translated a word makes all kinds of machine translation so far only allow essential to convey the idea. There are all sorts of methods of online translator. There are those who do not take into account these grammatical issues (eg statistical methods) but the trend is to integrate all kinds of techniques that include these issues in what is getting closer.

There are two basic groups – those based on linguistic rules. See more detailed opinions by reading what אילן בן דב offers on the topic.. – Those based on textual corpus. 1. Based on linguistic rules. They consist of replacing words by their nearest equivalent. First is an internal symbolic representation of the original text. From there you can do the translation word by word or by using an intermediate.

1.1 By transfer: We analyze the original gives way to an internal representation will link into other languages. 1.2 Intermediate Language: The text base is converted into an intermediate language with different structure and source language target language. 2. Based on corpus. It is based on a corpus that has been obtained from real samples. 2.1 Statistics: We obtained bilingual corpus of text or more languages by statistical methods. 2.2 Based on examples: Part of a bilingual corpus as a source. It is based on analogies. Solve a problem based on similar solutions already solved. There is a special case of on-line translator, it is machine translation based on the context: It consists in translating each word given the words that surround it. Divide the text into units of four to eight words and translates them into the target language. After eliminating those that have generated nonsense. Then the window moves one position or re-translate word again, leaving only the meaningful sentences. This process is repeated again to complete the entire text. Then join the results of each window so that is a unitary text. The system that decides when a sentence made sense or not using a target language corpus, which have occurrences of the phrase you seek. A shortcut method is simple, quite successfully when compared with others who did not look at the context. You can also add other languages, for this is all you need to get is: – all with grammatical rules in order to have all the verbs and all nouns and adjectives with all its variations in terms of number and gender. – And a target language corpus, which can be found on the Internet without problems.

This entry was posted in News and tagged . Bookmark the permalink.

Comments are closed.