Any idea on how to make a translator for unrecognised language.
Here in India we have many languages and I am planning/want to built my own language(mother tongue) translator that can translate from my language to english and vice versa like Google translator.
So any idea on how to do it. Do i need machine learning or do i need to play with just ascii/unicode only?
What do i need to learn or what programming language would be most suitable or any other courses?
Any suggestions would be appreciated.
Language translation makes sense if the languages have different writing scripts which basically means different alphabet set, few alphabets can be common but they are essentially different. Then it comes to their meaning and their usage which produce different results. If your language uses english alphabets then there is no point of translation, what you are probably looking for is equivalent semantic meaning. If you translate a sentence in English to another language then there are two things, first is translation of words to other language using the alphabets of other languages, and then the whole translation should make equivalent semantic sense.
To make semantic sense you will need dictionary lookup as well as some form of AI (machine learning).
And as said earlier you cannot translate an unrecognized language.
~ swim ~ l am from North-East India and i thought that I can contribute something to my clan via technology i thought of making a full working translator like google which translates a language sentence to english(eg hindi sentences to english) and just for information our language uses all the english alphabets.
"Continuing its stride, Google Translate has added support for seven more Indian languages on its app. These languages include Bengali, Gujarati, Kannada, Marathi, Tamil, Telugu, and Urdu."
only above languages are known by google translator
if this is a web application, then think about the place where the dictionary is saved (database), where it should be located, and the appearance of your translator. if you want to use ready-made dictionaries, then learn the API how to connect to existing dictionaries. then you only need to develop the interface (appearance)
You need machine learning like most translator apps does.
translating a language can't be specifically programmed, taking for example
he gave me the book,
he handed the book over to me
are pointing to the same thing in different ways in same language
Sorry i only know it is difficult if you start from scratch. But google translator does convert and make sense, so go through it's documentation and see if they have made some API available for public use.
Play with Ascii-Unicode
You cannot make a translator for unknown languages.
Unicode Standard needs to support the code-points and dialects. Then the display font needs to support the respective glyphs(font shapes).
Fortunately many Indian languages are supported by Unicode Standards, some dialects may not be supported though.
Doing it from scratch is tedious, look for some free libraries, i am sure there must be few.
If it is a web-based project then i think you can use google api for that.