Udupa R (2004) An English-Hindi statistical machine translation system. In: Conference on machine learning and cybernetics, pp 11–14 In: Workshop on NER for South and South East Asian Languages, IJCNLPĪntony PJ et al (2011) Kernel method for English to Kannada transliteration. Gali K et al (2008) Aggregating machine learning and rule based heuristics for named entity recognition. Pierrehumbert J, Nair R, Implications of Hindi prosodic structure. Katre DS (2006) A position paper on cross cultural usability issues of bilingual (Hindi & English) mobile phones. Goto I et al (2003) Transliteration considering context information based on the maximum entropy method. Computer Science and Engineering, Thapar University, Patiala Sharma N (2011) English to Hindi machine translation system. It shows that the performance is sufficiently good. The system works on the machine learning tools IRSTLM as a language model, GIZA++ for word alignment and MOSES for decoding. This system takes input in Devanagari lipi (i.e., script) and transliterates it into Latin script of English.
This work aims to investigate the problem of transliterating names and residential addresses written in Devanagari script into Roman Script using phonology context of Hindi, Urdu and English. The mapping of words written in Devanagari in residential address should be as per the phonetics of Hindi language and similar to Urdu and English. The major problem is identifying the context of named entity from language point of view and to spell it in the correct context. Even though Hindi is the national language of India, English is the official language used all over the country, and thus, these identity proofs contain the address in both the languages. This paper proposes to design and develop a system to give the correct spelling of the names and residential addresses which are mostly having named entities from different languages. which are our important identity proofs due to different context of making the spellings in Hindi, Urdu and English languages. It is observed that there are many spelling mistakes in names and residential addresses on Aadhar card, PAN card, driving license, etc. The places of accommodation or the residential addresses mostly consist of the named entities from Hindi, Urdu and English languages. There are people from Hindu, Muslim, Sikh and Christian religions living in India. As India is a versatile diverse nation, there are people from various communities speaking various languages.