Paraphrasing In Punjabi Language Using Natural Language Processing

Abstract

newline Till date, most of the automatic work in Natural Language Processing has newlinebeen done in English language but not much research work has been carried out in newlineIndian languages. One of such works is to convert the input given sentences in newlinedifferent forms by keeping the semantics or meaning same. This will be helpful in newlineconverting the complex sentences into simpler ones. In Natural Language Processing, newlinethe technical term used for such task is Paraphrasing. This system developed by us newlinemainly deals with Punjabi Sentences and its different forms. It takes a sentence as newlineinput and produces another sentence without changing its semantics. newlineIn this research work, we have presented the design and implementation newlinedetails of the paraphrasing system for Punjabi language along with its individual newlinecomponents, namely, sentence identification and classification system, sentence newlinesplitting system, clause conversion system, synonym-antonym replacement system newlineand number to text conversion system. The developed system can paraphrase Punjabi newlinesentences by splitting the long sentences into shorter ones and replacing the words in newlinethe sentence with their synonyms. It can paraphrase simple, compound and complex newlinesentences into their paraphrased forms. The paraphrased forms developed by this newlinesystem have same semantic meanings but may have different syntaxes. newlineA morphological feature based as well as structure of sentence based sentence newlineidentification system has been developed to identify and classify the simple, newlinecompound and complex sentences in Punjabi text. Regular expressions have been newlineused to split compound sentences into their independent clauses. Further, newlinemorphological features have been used to split a complex sentence into its newlinecomponents i.e. dependent clause and independent clause.