Tuesday, May 25, 2010
Thursday, May 20, 2010
Detection of complex predicate in bengali sentence
Take english - bengali parallel corpus . Take a bengali to english dictionary . Now first find the meaning of light verb in dictionary . Get all the synonyms of the word . now try to find out the meanings in the english sentence . If there is a matching then then it is not a complex predicate . If not found , then take the meaning of the previous word (main word ) with the light verb , if there is a matching in the english sentence , then it is a complex predicate . In case of compound verb take only the main verb meaning , if there is a matching , then it is a compound verb . This simmilar work was done by IIT kanpur ..
Compound verb in bengali by Soma Paul
Soma Paul has done a great job in bengali compound verb analysis . She has drawn a conclusion that there are 16 light verbs in bengali in compound verbs . Her phd thesis was on bengali compound verb .
serial verb in bengali
when two verb occure one by one in a sentence , and these two verb keep their meaning in the sentence unaltered then we will call these verbs are serial verb . Such as "Chele ta keye gelo "
The boy ate and went . Here two verb keye gelo occure side by side and they keep their meaning unaltered . so they are serial verb ..
The boy ate and went . Here two verb keye gelo occure side by side and they keep their meaning unaltered . so they are serial verb ..
Wednesday, May 12, 2010
Machine Translation in Natural Language Processing
Machine Transaltion is an interesting field in NLP . There are several kind of machine translation . Corpus based machine translation ( Statistical Machine Translation , Example Based Machine Translation ) is the most interesting field . Moses , Giza ++ are the tools for statistical Machine Tranlsation . SRILM tool kit is another important tool in SMT . In SMT allignment is a big issue . Allignment means rearranging the sentences in proper order as in the target language . In bengali and in english , the subject ,verb, object order is different from one to one . IBM research lab (Brown et all ) has done a great job in SMT . They use bayes theorem and a lot of probability calculation ..
Word Sense Disambiguation
A word may have multiple meaning . Such as
I am going to bank to deposit a cheque .
This river bank is very nice .
In the first sentence , bank means a financial institution , and in the second sentence meaning is a river bank . Now in case of Machine Translation you have to recognise which senese of a word is used in a particular sentence . Then translate the word with its proper meaning . So word sense disambiguation is a big problem in Machine Tralsation . There are several methods to solve the word sense disambiguation . One is Lesk Algorithm . It uses golssess of a sense of a word in Wordnet . Wordnet is a big resource to WSD .
I am going to bank to deposit a cheque .
This river bank is very nice .
In the first sentence , bank means a financial institution , and in the second sentence meaning is a river bank . Now in case of Machine Translation you have to recognise which senese of a word is used in a particular sentence . Then translate the word with its proper meaning . So word sense disambiguation is a big problem in Machine Tralsation . There are several methods to solve the word sense disambiguation . One is Lesk Algorithm . It uses golssess of a sense of a word in Wordnet . Wordnet is a big resource to WSD .
Explanation Of compound verb and conjucnt verb
Compound Verb :
Meya ta hese uthlo .
Here hese uthlo is a compound verb , because hese is the main verb containing the main meaning of the complex predicate and uhtlo is a light verb which is semantically bleached , that means it losses its literal meaning which is to rise , and creates a new meaning with main verb hese which is bursts into laughter .
Ami take biswas kori .
here biswas kori is a complex predicate and is a conjunct verb . here biswas is a noun which means believe and kori is a light verb which literal meaning is to do , but here this complex predicate creates a new meaning to believe .
Meya ta hese uthlo .
Here hese uthlo is a compound verb , because hese is the main verb containing the main meaning of the complex predicate and uhtlo is a light verb which is semantically bleached , that means it losses its literal meaning which is to rise , and creates a new meaning with main verb hese which is bursts into laughter .
Ami take biswas kori .
here biswas kori is a complex predicate and is a conjunct verb . here biswas is a noun which means believe and kori is a light verb which literal meaning is to do , but here this complex predicate creates a new meaning to believe .
Compound Verb and Conjunct Verb in Bengali
In bengali , there are compound verb and conjunct verb . In compound verb there are two verbs , the first one is main verb , which carries the main meaning of the complex predicate and the second verb is light verb which is semantically bleached . That means the light verb losses its literally meaning and creates a new meaning with the main verb .
In conjunct verb , there are two element , the first one is either a noun , or adjective or adverb and the second one is a verb that is light verb . Similar to the compound verb , the light verb losses its meaning and altogether creates a new meaning with the first part .
Example of compound verb :
Meye ta hese uthlo .
bai ta hat theke pore gelo .
Example of Conjuct verb :
ami take biswas kori .
Ami take bhorsa kori ..
In conjunct verb , there are two element , the first one is either a noun , or adjective or adverb and the second one is a verb that is light verb . Similar to the compound verb , the light verb losses its meaning and altogether creates a new meaning with the first part .
Example of compound verb :
Meye ta hese uthlo .
bai ta hat theke pore gelo .
Example of Conjuct verb :
ami take biswas kori .
Ami take bhorsa kori ..
Subscribe to:
Posts (Atom)
