rule based machine translation

Hi friend today let us know about rule based machine translation.this one of  approaches of machine translation and this is  earliest machine translation approach

Rule-based machine translation (RbMT) refers to a machine translation engine built on algorithms that analyze the syntax of the source language and uses rules to transfer the meaning to the target language by building a sentence. 

The rule based machine translation systems are developed  in early 1970’s they are

-Systran

-Japanese mt

-eurotra

At present following companies using  rule based machine translation are

-Apertium

-GramTrans

 The RBMT is based on linking structure  of given input with structure of demanded output service necessarily preserving their unique meaning.

For this process we need following

-          A dictionary that will map every world of source language to target language

-          Rules representing regular source sentence structure

-          Rules representing  target  sentence structure

The rgbt system contains following

·          SL morphological analyzer - analyses a source language word and provides the morphological information;

·          SL parser - is a syntax analyzer which analyses source language sentences;

·          translator - used to translate a source language word into the target language;

·          TL morphological generator - works as a generator of appropriate target language words for the given grammatical information;

·          TL parser - works as a composer of suitable target language sentences;

·         Several dictionaries - more specifically a minimum of three dictionaries:

a SL dictionary - needed by the source language morphological analyser for morphological analysis,

a bilingual dictionary - used by the translator to translate source language words into target language words,

a TL dictionary - needed by the target language morphological generator to generate target language words.

Advantages

-No bilingual texts required. Can be used for  translation that don’t have any  no texts in common

-Domain independent .rules will work on most of domains

-No quality ceiling. Every error can be corrected with targeted rule even if triggered case is rare

-Total control. Easy to debug since most of rules are hand written

- Reusability.  the source language analysis and target language generation parts can be shared between multiple translation systems, requiring only the transfer step to be specialized.

 Disadvantages

-          Unavailability of really  good dictionaries

-          Some linguistic information still needs to be set manually.

-          It is hard to deal with rule interactions in big systems, ambiguity, and idiomatic expressions.

-           Failure to adopt new domains