A new attempt to solve the problem of Hindi phonetics leveraging the Double Metaphone Algorithm.
Develop a library, which can match words based on similar phonetics for Hindi Language. Phonetics for example “डीपेश” & “दीपेश”.
In multilingual natural language processing system, there are many challenges in information retrieval. One of the biggest challenge is matching two set of strings which are not exactly same but are equivalent. For languages like Hindi where it is required to handle various pronunciations by people from various communities for the same keyword, especially in rural areas. This requires some sort of approximate matching algorithms (phonetic or semantic) to improve the recall of information retrieval systems. There are many phonetic algorithms available for English and other European languages, however for many Indian languages like Hindi they do not perform well. Having the matras in Hindi also adds up to the complexities where a small variation in spelling can impact the results heavily.
* Nishant Kumar
* Priyam Kumar
* Shushant Kumar
https://findwise.com/blog/things-to-consider-when-implementing-phonetic-search/