How it works!

Romtrans provides a general-purpose package for script to script conversion specifically Roman to non-roman scripts such as Devanagri, Arabic, Malayalam etc. It uses ID3 algorithm which is a machine learning algorithm for classification tasks.

Very simply, ID3 builds a decision tree from a fixed set of examples. The resulting tree is used to classify future samples. A training set in form of dictionary is provided to ID3 algorithm to make descision tree in the form of If-Else code.

Decision Tree is the representation of decision procedure for determining class (classification) of a given instance.Since roman script only has 26 characters whereas scripts such as devanagri have lot more characters to represent various sounds therefore such a classification is required.

The user must understand that script transliteration is not translation. Rather, script transliteration it is the conversion of letters from one script to another without translating the underlying words.

Using this desicion tree the input text in roman script is mapped to there corrseponding Phonemes in 7-bit ASCII characters(ISO 646 character set).For Roman to Hindi we have used ITANS (with some modifications) convention for representing phonemes.

Now these phonemes are mapped to unicode characters of the object language which is the final output.