Method and system for adaptive transliteration

• US 8,655,643 B2
• Filed: 10/09/2008
• Issued: 02/18/2014
• Est. Priority Date: 10/09/2007
• Status: Active Grant

• Current Assignee: Language Analytics, LLC
• Original Assignee: Language Analytics, LLC
• Inventors: Haddad, Habib, Jureidini, Imad
• Primary Examiner(s): COLUCCI, MICHAEL C
• Application Number:
  US12/248,422
• Publication Number:
  US 20090144049A1
• Time in Patent Office: 1,958 Days
• Field of Search: 704/5, 704/9, 704/8, 704/3, 704/277, 704/275, 704/260, 704/2, 707/999.004, 715/205, 709/218, 434/157, 370/352
• US Class Current: 704/3
• CPC Class Codes: G06F 40/129 Handling non-Latin characters, e.g. kana-to-kanji conversion

Abstract

A system and method for transliteration between two different character-based languages is provided. In some embodiments, the system and method provide transliteration from the Arabic language into Roman-based languages such as English. In some embodiments this system and method allows a user to more easily produce Arabic text on English or Roman-based computer hardware and software.

Claims

1. A computer-based method for adaptive transliteration between a first and a second character set, comprising:

receiving a set of input tokens in a first character set, entered into an input element of a user interface;

transliterating, in a processor, a subset of said set of input tokens substantially in real-time as said input tokens are received in said input portion of a user interface by comparing said subset of the input tokens against a database of known tokens in a second character set;

determining a set of output tokens in said second character set; and

providing a machine-readable output representative of said output tokens substantially in real-time responsive to receiving said input and upon processing and determining said output tokens.

2. The method of claim 1, said input tokens comprising a plurality of characters from said first character set.

3. The method of claim 1, said output tokens comprising a plurality of characters from said second character set.

4. The method of claim 1, said comparing step comprising comparing the input tokens with a corresponding group of possible entries in a database associating said input and said output tokens.

5. The method of claim 1, said input element comprising a text box adapted for receiving said input tokens.

6. The method of claim 1, said providing an output comprising providing an output to an output element of said user interface.

7. The method of claim 6, further comprising accepting a user selection from a plurality of possible options available to said user from said output.

8. The method of claim 7, further comprising processing said user selection and providing a result of said processing to affect future transliteration operations.

9. A system for adaptive transliteration between a first character set and a second character set, comprising:

a user interface input element for sequentially receiving input token entries in a first character set;

a processor for transliterating a subset of said input tokens substantially in real-time as they are received in said input element; and

an output element of said user interface for providing an output comprising at least one output token in said second character set, corresponding to said input tokens, substantially in real-time as the corresponding input tokens are received.

10. The system of claim 9, further comprising a database, coupled to said processor, for storing tokens in said first and second characters sets.

11. The system of claim 9, said input element comprising a text box for receiving said input tokens from a user.

12. The system of claim 9, said output element comprising an area for display of said output tokens to a user.

13. The system of claim 9, further comprising a transliteration score engine that receives a pair of associated tokens and provides an output indicative of a quantitative measure of the correlation between said pair of associated tokens.

14. The system of claim 9, further comprising a transliteration popularity database that receives a possible transliteration input and provides an output dependent on a metric indicative of the popularity of said possible transliteration.

15. A computer-based method for transliterating information, comprising:

sequentially receiving a set of input characters in a first character set in an input element of a user interface;

converting said input set of characters in said first character set into a set of input tokens in said first character set;

determining, in a processor, a plurality of transliteration matches to said input tokens from a possible set of output tokens in a second character set;

scoring, in said processor, said at least one match to determine a best match between said input tokens and said output tokens; and

presenting said output tokens to a user in an output element of said user interface, substantially in real time, based on said scoring such that a best suggested output token is preferentially presented along with other possible matches; and

analyzing a user selection of one of a plurality of presented output tokens.

16. The method of claim 15, said preferentially presenting of said best suggested output token further comprising sorting said output tokens so that said best suggested output token takes a primary place in said sorting.

17. The method of claim 15, further comprising storing a result of said user selection in a database for improving future transliteration operation.