"Turkish-Predictive-Text" project is a T9-styled Text Input Technology for Turkish which can be used in mobile phones or such devices.

It uses Zemberek NLP library as a validator for Turkish language.

The project is in alpha stage. That means it is still developing but if you insist of using it, it is at your own risk, there is no warranty.

Why Turkish-Predictive-Text?

• Mobile phone manufacturers do not support Turkish with T9™ very well.
• Contribution to open source Turkish NLP project repository.
• For fun ;)

TODOs

• Lookup table generation with all root+subfix combinations
• Word frequency ordering on alternative suggestions.
• Bug fixes and user-friendly GUI on Demo.

You can see Turkish-Predictive-Text demo at http://turkish-predictive-text.appspot.com/ui.jsp

Please checkout the code from SVN and if you want to contribute the project, and don't hesitate to drop an email to ahmet(at)ahmetalpbalkan.com

To checkout the code anonymously via Subversion (read-only):

svn checkout http://turkish-predictive-text.googlecode.com/svn/trunk/ turkish-predictive-text

Installation: Download last revision of turkish-predictive-text-*.jar and zemberek_tpt.jar, then add zemberek_tpt package to your CLASSPATH.

NOTE: This project is not affiliated with T9™ and Nuance Communications Inc.

NOTE: Zemberek is distributed under Mozilla Public License 1.1.

Current maintainer: ahmet alp balkan <ahmetalpbalkan at gmail.com> Please do not hesitate to contact for contributing the project.