The project aims to provide a set of tools for extracting cognates and false friends from text corpora. Currently it supports Bulgarian, Russian and partially English.

More information: http://www.nakov.com/blog/2009/09/30/open-source-toolkit-for-extraction-of-cognates-and-false-friends-tecff/

Implemented Algorithms

The toolkit provides implementation of the following algorithms:

• MMEDR: measures modified orthographic similarity between Bulgarian and Russian words
• SemSim: measures semantic similarity betwen words by searching in Google and analysing the returned text snippets (supports Bulgarian, Russian and English)
• CrossSim: measures cross-lingual semantic similarity by searching in Google and analysing the returned text snippets (supports Bulgarian and Russian)
• FFExtract: extracts false friends from parallel corpus by determining candidates through MMEDR algorithm and combining statistical and semantic evidence for distinguishing between cognates and false friends

Source Code

The source code is written in C# for .NET Framework 3.5 and can be compiled with Visual Studio 2008 Express Edition. The source code is available from its public SVN repository: http://cognates-and-false-friends-tools.googlecode.com/svn/trunk/.

Project Roadmap

The project is still in early alpha version.